Empowered Parents

10 Hands-On Shapes Recognition Activities for Preschoolers

By: Author Tanja McIlroy

Posted on Last updated: 25 January 2023

Categories Early Mathematical Skills

shape based activity recognition

Teaching children about shapes should be a fun, hands-on experience. It is by experiencing concepts through the body and the senses that real learning happens in the early years.

Scroll down to the section ‘10 simple shapes recognition activities’ for some great ideas.

Why is it Important to Learn About Shapes?

An important part of child development, shape recognition is about more than just being able to spot the basic shapes. It’s important for building early maths , reading and other areas too.

Shapes are found all around us and take many different forms, both two-dimensional and three-dimensional.

Children learn that shapes have properties – such as lines, curves, sizes, lengths and positions in space , and that all physical objects have shape. 

They develop their visual perception (when the brain interprets what the eyes see) as they are exposed to shapes around them. 

Shapes are a part of mathematical concepts (like geometry) and are also used in many scientific and technical fields.

The recognition of shapes is crucial for learning to read and building early literacy skills . 

Letters are made up of shapes and children need to learn to distinguish, recognize and remember the specific shapes in order to read fluently. Shapes are the first ‘symbols’ children learn to interpret. [ source ]

Child making shapes with blocks

How Do You Teach Shape Recognition?

The best way for children to learn about shapes is through play. They need many opportunities to touch and see three-dimensional shapes in their daily lives.

These concepts can also be taught through shape recognition games and activities that are planned specifically to introduce the concept of shape.

Activities like tracing and worksheets are best kept for much older children as these are unlikely to hold a child’s attention for long, and they are not really meaningful to a young child.

It is more important to focus on doing hands-on activities during preschool so that a child fully understands the properties of shapes by experiencing them with their body and all their senses .

It is also important for parents and teachers to actively teach the vocabulary about shapes. They need to hear concepts such as:

  • Straight lines or curvy lines
  • Size – bigger, smaller, longer, etc.
  • Round, pointy
  • 2-dimensional, 3-dimensional
  • The names of shapes – oval, rectangle, triangular prism

This can be done in natural ways during a conversation:

  • Is your dinner plate round? 
  • What else is round in our house?
  • This picture of a tent looks like a triangle – it’s pointy at the top.

When teaching the shapes, remember children learn these in a concrete-pictorial-abstract way, as they do many other concepts. 

They learn the shape of a physical object, and then later learn to associate a picture of a shape with the physical object. After that, they can recognize an abstract representation of a shape.

An example of this is playing with a beach ball and learning that it is round; then understanding that a picture of a ball represents a round, circular shape; then identifying a simple circle drawn on paper.

10 Simple Shapes Recognition Activities and Games

Here are some simple shape activities to try at home or school.

Blocks lying on a table. Text reads

1. Shape Sorting

Sorting is an important mental concept and it’s easy to do with shapes.

Provide a container of shapes – cutouts, wooden or plastic shapes – and get kids to sort them according to different criteria. 

Here are some examples:

  • Shapes with straight edges vs shapes with curved edges
  • Shapes with 3 sides and shapes with 4 sides
  • Shapes with corners and shapes without corners
  • 2D vs 3D shapes

Here are more fun sorting activities for preschoolers .

2. Body Shapes

Play a movement game with shapes. Say the name of a shape and ask children to make the shape with their bodies. 

Get them to straighten up into a tall rectangle, puff their arms out to the side to make themselves round, stand with legs far apart and arms pointed above you to make a triangle, etc.

Before you show or help with ideas, see if the kids can figure out on their own how to make these shapes. They may think outside the box and find new ways to make the shapes. 

This is a great activity for body awareness .

3. Block and Ball Play

Children need regular exposure to shapes by playing with blocks, balls and other toys during their free play time.

This is an activity that doesn’t need much regulation by an adult but is as important as planned shapes activities.

4. Playdough Shapes

Give kids playdough and shape cutters and let them explore different shapes this way.

Make a batch of easy homemade playdough and this will keep kids occupied for hours.

If you don’t have cutters in the form of shapes, then improvise with tins (round), plastic containers (square and rectangle), etc.

Preschooler playing with playdough

5. Yummy Shapes

For a tastier version of the above activity – that will leave a bigger imprint on kids’ memory – make shapes out of cookie dough and then eat them after!

6. Find the Shapes

Show children an example of a simple shape, such as a circle, and challenge them to find as many items as they can around them that have the same shape. (Here are more games about circles to try).

They could also draw pictures of all the things they can think of that have a circle shape.

This teaches kids that shapes are part of their world.

7. Build With Shapes

If you have plastic or wooden shapes that are relatively flat, let kids ‘build’ pictures out of them using the shapes. 

Children playing with tangram shapes

They could construct a house or a person, for example, by learning to use the shapes to substitute the various parts they need.

These shapes can also be made out of cardboard.

8. Feel the Shape

Use the same shapes from the activity above to do a sensory activity.

Blindfold a child, then give him a shape or ask him to pick a shape out of a bag. He must describe the shape (it is round and has no straight edges), then say what the shape is. 

If he guesses right, he can keep the shape. If not, it goes back into the bag. This can be turned into a game to see how many shapes each person can accumulate.

9. Shape Stamping

Provide poster paints on trays and get kids to stamp shapes onto them.

Use blocks for these shapes or make potato prints , sponge shapes, etc. Use your imagination.

10. Shape Pictures

A fun art activity, this can be attempted after children have had ample opportunity to make shape pictures out of real shapes. 

Provide lots of pre-cut paper shapes in various colours, white paper and glue.

Get kids to make their own shape pictures. Let them use their creativity and do not try to guide their creations. 

Definitely don’t give them a model to copy! This should be a form of process art , where the learning is in the doing, not the final product.

If you are looking for more creative art ideas, try these shape crafts for preschoolers .

“ Learning Through Play: A parent’s guide to the first five years ” , written by Jan Natanson.

“ Language and School Readiness ” , written by Martie Pieterse.

Get FREE access to Printable Puzzles, Stories, Activity Packs and more!

Sign up and you’ll receive a downloadable set of printable puzzles, games and short stories , as well as the Learning Through Play Activity Pack which includes an entire year of activities for 3 to 6-year-olds. Access is free forever.

Signing up for a free Grow account is fast and easy and will allow you to bookmark articles to read later, on this website as well as many websites worldwide that use Grow .

Printables and Learning Through Play Activity Pack

This site uses Akismet to reduce spam. Learn how your comment data is processed .

Taming Little Monsters

25 Shape Activities for Kids

Categories Activities

All of these Shape Activities for Kids introduce toddlers and preschoolers to shapes through play-based learning. 25 ways to recognise, create, and explore basic shapes.

If you want to make your Shapes themed lessons a breeze, then check out the Shapes Activity Pack . It’s filled with math and literacy centers, fine motor activities and arts and crafts templates. All of which are aligned with preschool learning standards. Check it out today.

shape based activity recognition

Disclosure: Some of the links provided in this blog are affiliate links. I will be paid a commission if you use this link to make a purchase.

Shape Arts and Crafts for Kids

1. block painting.

shape based activity recognition

Blocks come in a variety of shapes. Circles, triangles, squares. Any picture can be broken down into these basic shapes. If your little ones are too young to make an actual picture, don’t worry, it’s still a lot of fun dunking the blocks in the paint and making shapes on a page.

2. Giant Shape Art for Kids

Giant shape art for kids. A fun way for toddlers and preschoolers to learn shapes. Include it in your list of shape activities for kids.

Giant Shape Art for Kids was a fun way to spend an afternoon with the kids. We used an old cardboard box and painters tape to create some fun tape resist art, exploring and learning all about shapes as we did it. If you’re teaching a toddler or preschool classroom, try this simple art activity, the kids will love it.

3. Black Glue Shape Art for Kids

Learn how to make Black Glue Shape Art for Kids. A fun and easy kids art idea for toddlers and preschoolers who are learning all about shapes

Black Glue Shape Art for Kids is all about learning 2D shapes. Toddlers and preschoolers love art ideas for kids that are fun, while parents and teachers love art ideas that are easy and educational. This activity ticks all of those boxes.

Either make single pages for home, or make large ones as a small group project in the classroom. This activity can easily be adapted for both parents and teachers.

4. Kandinsky Shape Art for Kids – Messy Little Monster

shape based activity recognition

Explore colors and shapes with your toddler or preschooler with this shape art for kids. This hands on learning art project inspired by Kandinsky. Is fantastic for kids of all ages.

5. Recycled Shapes Process Art – Mosswood Connections

shape based activity recognition

Raid the recycle bin and use recycled items to make art. Kids will have fun discovering shapes with this Process Art Project.

Shape Sensory Activities for Kids

6. shape sorting sensory bin.

shape based activity recognition

This shape sorting sensory bin is loads of fun for toddlers and preschoolers. Get the FREE printable shapes and start teaching some math.

7. Shape Match Sensory Bin – Happy Toddler Playtime

shape based activity recognition

Shape Match Sensory Bin is a fun sensory bin using shredded paper. It’s a fun and easy way to learn and practice shape recognition with your toddler or preschooler.

8. Fizzy Shapes Sensory Bin – Happy Toddler Playtime

shape based activity recognition

Shapes play ideas where science meets math will always be a hit with toddlers and preschoolers. Check out this fizzy shapes sensory bin.

9. Shape Button Sensory Bin – 3 Dinosaurs

shape based activity recognition

Have you seen those buttons that are different shapes? If not, go and grab them because you can create so many different play based learning activities with them.

10. Shape Sensory Squish Bag – Still Playing School

shape based activity recognition

Create a sensory squish bag for kids to learn shapes. Stick it on the window to see what happens when the sun shines through.

Shape Science Activities

11. building shapes – stem for kids.

Building shapes STEM for kids is a fun and easy engineering activity for kids. Help toddlers and preschoolers learn their shapes with some play based learning.

This Building Shapes – STEM for Kids activity is a fun, play based learning activity for toddlers and preschoolers. Engineering for kids can be both easy and fun, and includes more versatility than just building blocks.

12. Which Shape is Strongest – Science Sparks

shape based activity recognition

Did you know that under certain circumstances paper can be very strong? This easy investigation uses paper folded into different shaped columns to hold up books. Which shape will hold the most books, before crumbling under the weight?

13. 3D Bubble Shape Activity – Little Bins for Little Hands

shape based activity recognition

Can you make bubbles different shapes? This is one of my favorite shape activities for kids because you find out how to make your own cool 3D bubble wands and explore the science of bubbles.

14. Floating Shapes Experiment – Active Littles

shape based activity recognition

Want to learn how to make the floating dry erase marker experiment work? Learn how to make shapes float around on the top of water, just like magic.

15. Magic Disappearing Shapes Activity – Gift of Curiosity

shape based activity recognition

This simple shapes activity brings some magic to your child’s learning by disappearing at your child’s command. Not bad for paper towels and water, is it?

Shape Fine Motor Activities for Kids

16. sticker shapes fine motor activity.

This sticker shapes fine motor activity is a fun way for toddlers and preschoolers to learn shapes and develop fine motor skills.

This sticker shapes fine motor activity is a fun way for toddlers and preschoolers to learn shapes and develop fine motor skills.

17. Mystery Shapes – Days of Grey

shape based activity recognition

Mystery Shapes is a fun fine motor activity for toddlers and preschoolers. Join the dots together to discover what shapes are hidden there.

18. Basic Shapes Work Station – Learning 4 Kids

shape based activity recognition

Shapes play ideas provide children with the opportunity to explore and create with shapes. It teaches shape recognition and has the benefit of developing fine motor skills.

19. Sand Shapes – Gift of Curiosity

shape based activity recognition

Sand shapes, a shape themed art, and fine motor activity kids will love.

20. Playdough Shapes Puzzle – Mama OT

shape based activity recognition

Help your child develop fine motor skills with this simple yet creative puzzle activity using shapes and play dough.

Shapes Free Printable Activities

21. craft stick shape mats.

Get your copy of these FREE Craft Stick Shape Mats. Use them to teach toddlers and preschoolers math, fine motor skills, language and more.

Get your copy of these FREE Craft Stick Shape Mats. They’re perfect to help toddlers and preschoolers who are learning about shapes.  The children will develop fine motor skills, math, language, and more with this easy, low-prep activity.

22. Shape Playdough Mats – Fantastic Fun and Learning

shape based activity recognition

Use these shape play dough mats to help preschoolers and kindergarteners learn how to make 2D shapes and recognize shapes in everyday objects

23. Free Shape Lacing Cards – 3 Dinosaurs

shape based activity recognition

When you’re looking for shape activities for kids, make sure you include some shape lacing cards. This is a fun way to learn shapes with some hand-on learning.

24. Free Shape Flashcards – Look We’re Learning

shape based activity recognition

These free printable shape flashcards are a wonderful learning resource for toddlers, and preschoolers who are learning about 2-D shapes.

25. Free Shape Dot Marker Printables – 3 Dinosaurs

shape based activity recognition

Use this printable with dot markers, playdough or manipulatives. This simple printable is so incredibly versatile, your students will love it.

26. Shape Stamp Painting

Try this shape stamp painting for kids. A fun and easy shape art and craft idea for toddlers and preschoolers.

Make your own shape painting using shape stamps . This is an asy way to use those shape sorting toys in a fun new way.

Are you going to try any of these Shape Activities for kids? Don’t forget to pin the idea for later.

shape based activity recognition

More Play Based Learning Activities

shape based activity recognition

Shapes Activity Pack

25 Hands on alphabet activities for kids. Learn the ABC's with these play based learning ideas for toddlers, preschoolers and kindergarte,

25 Alphabet Activities for Kids

25 Weather activities for kids. Weather themed play based learning ideas for toddlers and preschoolers. Including science, fine motor, sensory, free printables and arts and crafts ideas.

25 Weather Activities for Kids

Share this:

  • Click to share on Facebook (Opens in new window)
  • Click to share on Pinterest (Opens in new window)
  • Click to share on Twitter (Opens in new window)
  • Skip to main content
  • Skip to primary sidebar

Teaching 2 and 3 Year Olds

This post might contain affiliate links. Click here for more information . Thanks for visiting!

How to Teach Shape Recognition to Preschoolers with Fun Activities

September 13, 2017 by Sheryl Cooper

Inside: How to Teach Shape Recognition to Preschoolers with Fun Activities with Free Printable

One of the areas we have to cover in our preschool assessments is shapes. I am always looking for ways to build shape recognition skills, but I want them to be playful and authentically blend in with our other hands-on activities.

I decided to put together my own collection of shape recognition activities by doing a little searching.

After weeding out the results, I came up with over 16 ideas that I knew would be perfect. Not only are they fun, but they involve building other skills , too. I always love that!

So put away those flashcards, my friends. These fun activities will help your preschoolers learn to identify circles, triangles, squares, rectangles, and ovals .

👉 Bonus! I created a shape matching activity printable for your children. Scroll towards the bottom of this post to get the download.

  • Preschool Math Activities
  • Preschool Literacy Activities

How to Teach Shape Recognition to Preschoolers with Fun Activities

Build shape recognition using toys .

Make circles by painting with spools .

Play a shape game while also recognizing colors.

Make circles with a DIY circle stamper .

preschool shape activities

Mix colors while stamping circles from paper tubes .

Take apart your nesting toys to stamp shapes with paint .

Make a simple shape collage . (Homegrown Friends)

Go on a shape hunt at the playground (free printable included). (Buggy and Buddy)

preschool shape activities

Paint with different shapes of foam blocks. (Meri Cherry)

Make a shape puzzle using blocks you already have. (Twodaloo) (ETA: Unfortunately website is no longer running.)

Trace shapes on a chalkboard. (Hands On As We Grow)

Create a fine motor shape bin . (Stir the Wonder)

preschool shape activities

Make a shape sorting activity from an ice cube tray. (Mess for Less)

Make paper roads in different shapes and trace with cars. (Adventures of Adam)

Play shape hopscotch indoors! (Housing a Forest)

Use colored construction paper to create a train color sorting activity . (Mom Inspired Life)

preschool shape activities

More shapes activities:

Toddler Triangle Shapes Activity

15 Activities for Teaching Squares

Shapes Art Using Toys

Teaching Circles to Toddlers and Preschoolers

Going on a Circle Hunt Circle Time Activity

Here’s a simple shape matching activity for your children to try today! Click on the photo for the link to the PDF download:

How to Teach Shape Recognition to Preschoolers with Fun Activities

FREE CIRCLE TIME PLANNER!

Get your FREE circle time planner as a gift when you subscribe to my free weekly newsletters.

Here is my Privacy Policy

Success! Now check your email to confirm your subscription.

There was an error submitting your subscription. Please try again.

' src=

About Sheryl Cooper

Sheryl Cooper is the founder of Teaching 2 and 3 Year Olds, a website full of activities for toddlers and preschoolers. She has been teaching this age group for over 20 years and loves to share her passion with teachers, parents, grandparents, and anyone with young children in their lives.

Reader Interactions

' src=

May 31, 2016 at 5:32 pm

These are great ways to teach shapes and colors. I love how you said in your intro that kids don’t need worksheets or flashcards. I’m a firm believer in hands on learning, especially for our little ones.

' src=

May 31, 2016 at 10:46 pm

Thank you, Amanda! I feel strongly about this. Thank you for reading!

' src=

September 14, 2017 at 8:39 am

Thanks so much sheiyl for this post I really love it as a preschooler teacher.

September 14, 2017 at 9:00 pm

I’m happy to hear that! Thank you!

  • Grades 6-12
  • School Leaders

☘️ St. Patrick's Day Activities: Books, art ideas, experiments, and more!

25 Creative Activities and Ideas For Learning Shapes

Triangles, circles, and squares… oh, my!

Five separate images of activities to learn shapes from paper to rocks.

Learning shapes is one of the earliest concepts we teach kids. This readies them for geometry in the years ahead, but it’s also an important skill for learning how to write and draw. We’ve rounded up our favorite activities for learning shapes, both 2-D and 3-D. They all work well in the classroom or at home.

1. Start with an anchor chart

shape based activity recognition

Colorful anchor charts like these are terrific reference tools for kids learning shapes. Have kids help you come up with examples for each one.

Learn more: A Spoonful of Learning / Kindergarten Kindergarten

2. Sort items by shape

shape based activity recognition

Collect items from around the classroom or house, then sort them by their shapes. This is a fun way for kids to realize that the world around them is full of circles, squares, triangles, and more.

Learn more: Busy Toddler/Shape-Sorting

3. Snack on some shapes

Learning Shapes Chieu Urban

Everyone loves a learning activity you can eat! Some food items are already the perfect shape; for others, you’ll have to get a little creative.

Learn more: Chieu Anh Urban

[contextly_auto_sidebar]

4. Print with shape blocks

Learning Shapes Pocket of Preschool 2

Grab your shape blocks and some washable paint, then stamp shapes to form a design or picture.

Learn more: Pocket of Preschool

5. Go on a shape hunt

Learning Shapes Nurture Store

These “magnifying glasses” make an adventure of learning shapes! Tip: Laminate them for long-term use.

Learn more: Nurture Store UK

6. Hop along a shape maze

shape based activity recognition

Use sidewalk chalk to lay out a shape maze on the playground or driveway. Choose a shape and hop from one to the next, or call out a different shape for every jump!

Learn more: Creative Family Fun

7. Assemble a truck from shapes

Learning Shapes Little Family Fun

Cut out a variety of shapes (excellent scissors skills practice!), then assemble a series of trucks and other vehicles.

Learn more: Little Family Fun

8. Stretch out shapes on geoboards

shape based activity recognition

Teachers and kids love geoboards , and they’re a great tool for learning shapes. Give students example cards to follow, or ask them to figure out the method on their own.

Learn more: Mrs. Jones’ Creation Station

9. Drive on shaped roads

shape based activity recognition

Use these free printable road mats to work on shapes. Bonus: Make your own road shapes from sentence strips!

Learn more: PK Preschool Mom

10. Find shapes in nature

shape based activity recognition

Take your shape hunt outside and look for circles, rectangles, and more in nature. For another fun activity, gather items and use them to make shapes too.

11. Put together craft stick shapes

Learning Shapes Surviving a Teachers Salary

Add Velcro dots to the ends of wood craft sticks for quick and easy math toys. Write the names of each shape on the sticks for a self-correcting center activity.

Learn more: Surviving a Teacher’s Salary

12. Blow 3-D shape bubbles

shape based activity recognition

This is a STEM activity that’s sure to fascinate everyone. Make 3-D shapes from straws and pipe cleaners, then dip them in a bubble solution to create tensile bubbles. So cool!

Learn more: Babble Dabble Do

13. Prep a shape pizza

Learning Shapes Mrs Thompsons Treasures

Cover a paper plate “pizza” with lots of shape toppings, then count the number of each. Simple, but lots of fun and very effective.

Learn more: Mrs. Thompson’s Treasures

14. Construct shapes from toothpicks and Play-Doh

shape based activity recognition

This is an excellent STEM challenge: how many shapes can you make using toothpicks and Play-Doh? Marshmallows work well for this activity too.

Learn more: Childhood 101

15. Outline shapes with stickers

Learning Shapes Busy Toddler 2

Kids adore stickers, so they’ll enjoy filling in the outlines of the shapes they’re learning. They won’t realize it, but this gives them fine motor skills practice too!

Learn more: Busy Toddler/Sticker Shapes

16. Lace shapes

Learning Shapes Planning Playtime

Lacing cards have long been a classic, but we really like this version that uses drinking straws. Just cut them into pieces and glue them along the edges of the cards.

Learn more: Planning Playtime

17. Make shapes with LEGO bricks

Learning Shapes Pocket of Preschool

LEGO math is always a winner! This activity also makes a good STEM challenge. Can your students figure out how to make a circle from straight-sided blocks?

18. Categorize shapes by their attributes

Learning Shapes Susan Jones Teaching

Work on geometry terms like “sides” and “vertices” when you sort shapes using these attributes. Start by placing shapes into paper bags and asking students questions like, “The shape in this bag has 4 sides. What could it be?”

Learn more: Susan Jones Teaching

19. Count and graph shapes

Learning Shapes Playdough to Plato

These free printable worksheets challenge kids to identify shapes, then count and graph them. Lots of math skills, all in one!

Learn more: Playdough to Plato

20. Create a shape monster

Learning Shapes Fantastic Fun and Learning

Add arms, legs, and faces to create cheery (or scary) shape monsters! These make for a fun classroom display.

Learn more: Fantastic Fun and Learning

21. Sift through rice for shapes

shape based activity recognition

Sure, kids can identify their shapes by sight, but what about by touch? Bury blocks in a bowl of rice or sand, then have kids dig them out and guess the shape without seeing them first.

Learn more: Fun With Mama

22. Craft an ice cream cone

shape based activity recognition

Ice cream cones are made up of several shapes. Encourage kids to see how many different ways they can make a sphere of “ice cream.”

Learn more: Extremely Good Parenting

23. Ask “What does the shape say?”

Learning Shapes Around the Kampfire

If you don’t mind the risk of getting that song stuck in your kids’ heads, this is such a neat way to combine writing and math.

Learn more: Around the Kampfire

24. Piece together shape puzzles

Learning Shapes Toddler at Play

Use wood craft sticks to make simple puzzles for kids who are learning their shapes. These are inexpensive enough that you can make full sets for each of your students.

Learn more: Toddler at Play

25. Feed a shape monster

Learning Shapes Teach PreK

Turn paper bags into shape-eating monsters, then let kids fill their hungry bellies!

Learn more: Teach Pre-K

From teaching shapes to long division and everything in between, these are the 25 Must-Have Elementary Classroom Math Supplies You Can Count On .

Plus, 22 Active Math Games and Activities For Kids Who Love to Move .

25 Creative Activities and Ideas For Learning Shapes

You Might Also Like

Collage of First Grade Math Games, including Shape Guess Who? and Addition Tic-Tac-Toe

30 Fun and Free First Grade Math Games and Activities

Teach them early on that math can be fun! Continue Reading

Copyright © 2023. All rights reserved. 5335 Gate Parkway, Jacksonville, FL 32256

shape based activity recognition

  • Primary Hub
  • Art & Design
  • Design & Technology
  • Health & Wellbeing
  • Secondary Hub
  • Citizenship
  • Primary CPD
  • Secondary CPD
  • Book Awards
  • All Products
  • Primary Products
  • Secondary Products
  • School Trips
  • Trip Directory
  • Trips by Subject
  • Trips by Type
  • Trips by Region
  • Submit a Trip Venue

Trending stories

Actor playing Lady Macbeth

Top results

shape based activity recognition

8 of the best shape recognition resources and activities for early years

shape based activity recognition

Don't be a square, boost your kids' shape recognition skills with these ace activities

Teachwire

1 | We’re going on a shape hunt

shape based activity recognition

Grab your iPads and help children explore their environment, with these ideas from Marc Faulder.

Children will learn how to talk about the properties and functions of 2D shapes, recognise 2D shapes in the environment and use knowledge of 2D shapes in a design task.

You’ll find this free lesson plan here.

2 | 2D shape recognition posters

shape based activity recognition

This colourful set of 12 posters will introduce children in the early years and Key Stage 1 to common regular shapes, from the simple circle and square to the crescent and decagon.

Display them around your classroom to develop important mathematical language and knowledge.

Download these free printable posters here.

3 | Wooden shape puzzles

shape based activity recognition

For those of you who are handy with a hacksaw and some sandpaper these colourful and chunky shape puzzles can be made with minimal fuss.

All the instructions you need, should you need them, are here.

4 | Printable shape puzzles

shape based activity recognition

Alternatively, if scissors and glue are more your thing than sweeping up sawdust these printable puzzle pieces might be a better option.

You’ll find these here.

5 | Name that shape

shape based activity recognition

This colourful selection of illustrations are formed of a variety of shapes, designed to support the development of early maths skills.

Sheets containing the constituent parts of each picture are provided for you to cut up, allowing children to recreate the illustrations or come up with their own designs.

You can find all these here.

6 | Feed the shape monsters

shape based activity recognition

Give shape sorting a terrifying twist in time for Halloween with these googly-eyed shape monsters.

Head to The Imagination Tree for this fun sorting game.

7 | Cut-out road shapes

shape based activity recognition

These six cut-out road designs for early years settings will support children’s shape recognition and develop their motor skills as they engage in vehicle-based small-world play.

Print them out here.

8 | Mirror box shapes

shape based activity recognition

This mirror box idea is just one of many shape-based activities at Stimulating Learning .

You’ll also find some cool lightbox creations, shape picture activities and some doughy monster making.

Check them all out here.

Sign up to our newsletter

You'll also receive regular updates from Teachwire with free lesson plans, great new teaching ideas, offers and more. (You can unsubscribe at any time.)

Which sectors are you interested in?

Early Years

Thank you for signing up to our emails!

You might also be interested in...

shape based activity recognition

Why join Teachwire?

Get what you need to become a better teacher with unlimited access to exclusive free classroom resources and expert CPD downloads.

Exclusive classroom resource downloads

Free worksheets and lesson plans

CPD downloads, written by experts

Resource packs to supercharge your planning

Special web-only magazine editions

Educational podcasts & resources

Access to free literacy webinars

Newsletters and offers

Create free account

I would like to receive regular updates from Teachwire with free lesson plans, great new teaching ideas, offers and more. (You can unsubscribe at any time.)

By signing up you agree to our terms and conditions and privacy policy .

Already have an account? Log in here

Thanks, you're almost there

To help us show you teaching resources, downloads and more you’ll love, complete your profile below.

Welcome to Teachwire!

Set up your account.

Lorem ipsum dolor sit amet consectetur adipisicing elit. Commodi nulla quos inventore beatae tenetur.

Log in to Teachwire

Not registered with Teachwire? Sign up for free

Reset Password

Remembered your password? Login here

close

Get Your ALL ACCESS Shop Pass here →

Little bins for little hands logo

25 Shape Activities For Preschool and Kindergarten

Shapes are a fun and engaging topic for preschoolers, that provide an important foundation for future mathematical learning. Explore how shapes can be integrated into a STEM curriculum. From shapes in everyday life, to shape art activities and even science, check out these easy to set up shape activities for preschoolers and kindergarten. Plus, make sure to grab the printable shape activities pack at the end!

shape based activity recognition

Why Teach Preschoolers About Shapes?

Learning about shapes is important for preschoolers for several reasons:

Increases Their Vocab

Learning shapes introduces new words into a child’s vocabulary. Describing and talking about shapes helps preschoolers improve their language skills as they learn to express themselves more precisely.

Develops Pre-Math Skills

Recognizing shapes is an early introduction to mathematical concepts. It lays the groundwork for understanding concepts like counting, patterns, and spatial relationships, which are essential in math education.

Promotes Shapes In Art

Shapes are a fundamental component of art and creativity. Understanding shapes enables preschoolers to create more complex and visually appealing artwork.

Shapes In Everyday Life

Shapes are all around us in everyday objects. Learning about shapes helps children make connections between the abstract concept of shapes and the physical world they interact with.

Develops Problem Solving Skills

Understanding shapes also contributes to spatial awareness, which is crucial for tasks like reading maps, following directions, and understanding the arrangement of objects in space.

Manipulating shapes and fitting them into various puzzles or activities enhances a child’s problem-solving skills and encourages them to think creatively about spatial challenges.

Learning about shapes in preschool really does lay the groundwork for a wide range of cognitive, linguistic, mathematical, and creative skills. It is also foundational for later learning in science, technology, engineering, and mathematics (STEM) fields. Read more about STEM and teaching shapes further on.

Books About Shapes

Here are five shape books perfect for pairing with these shape activities! ( I am an Amazon Affiliate )

“Mouse Shapes” by Ellen Stoll Walsh : This interactive book follows three mice as they explore different shapes to escape from a cat. It’s a great way for preschoolers to learn shapes while engaging in a cute story.

“ Brown Rabbit’s Shapes ” by Alan Baker : Join Brown Rabbit as he discovers shapes in his everyday life, from squares to stars. The colorful illustrations make learning about shapes enjoyable.

“ Shape by Shape ” by Suse MacDonald : This book combines clever die-cuts with colorful illustrations to explore how shapes can be transformed into various objects and animals.

“ Color Zoo ” by Lois Ehlert : While not exclusively about shapes, this book introduces animals created from basic geometric shapes. The cut-paper illustrations are vibrant and engaging.

“ The Shape of My Heart ” by Mark Sperring and Alys Paterson : Follow Little Polar Bear as he explores his world and discovers shapes in unexpected places. The rhythmic text and charming illustrations make it a delightful read.

For more excellent suggestions for book about shapes, see our printable shape worksheet pack at the end!

List of 2D Shapes

Here’s a list of different shapes along with simple descriptions that you can use to teach kids about two dimensional or 2D shapes.

Circle : A circle is a round shape with no corners or edges. It looks like the outline of a ball or a plate.

Square : A square has four equal sides and four right angles. All its sides are the same length, and it looks like a box or a piece of paper.

Triangle : A triangle has three sides and three angles. It can have different types, like equilateral (all sides are the same length), isosceles (two sides are the same length), or scalene (all sides are different lengths).

Rectangle : A rectangle has four sides with opposite sides being equal in length and all angles being right angles. It looks like a stretched-out square.

Oval : An oval is like a stretched-out circle. It has curved sides but no corners or edges.

Hexagon : A hexagon has six sides and six angles. It’s like a stop sign with six equal sides.

Pentagon : A pentagon has five sides and five angles. It’s like a house with a slanted roof.

Star : A star has many points, and it can vary in the number of points it has. It can look like the shape you draw when you connect the dots of a star in the sky.

Heart : A heart shape is often associated with love. It has two rounded parts at the top and comes together at a point at the bottom.

Diamond/Rhombus : A diamond (also called a rhombus) has four equal sides, but its angles are not right angles. It looks like a tilted square.

Crescent : A crescent is a shape that looks like a curved moon. It’s like a circle with a chunk missing.

How To Teach Shapes With STEM

Teaching preschoolers about shapes can be directly connected to STEM (Science, Technology, Engineering, and Mathematics) education. Here are some ideas for how to teach shapes to young kids using the components of STEM education.

  • Encourage preschoolers to explore their environment and find objects that match specific shapes. This fosters observation skills and scientific inquiry.
  • Help children recognize patterns created by repeating shapes. Patterns are a fundamental concept in science and math.
  • Discuss shapes found in nature, such as the shapes of leaves, flowers, and animals. This connects shapes to the natural world and promotes curiosity.

Technology:

  • Use interactive apps or digital games that teach shapes through touchscreen activities. This introduces technology while enhancing shape recognition.
  • Some online platforms offer virtual manipulatives that allow kids to interact with shapes and build structures.

Engineering:

  • Provide building materials like blocks, straws, and connectors and let kids construct shapes and structures. This encourages hands-on engineering skills.
  • Present challenges like “build a tower using only triangular blocks” to engage children in engineering thinking.

Mathematics:

  • Shapes are a fundamental aspect of geometry. Use shape recognition games, puzzles, and activities to introduce basic geometric concepts.
  • Have children sort objects by shape, creating opportunities for classification and mathematical thinking.
  • Manipulating and combining shapes helps develop spatial awareness and understanding, which is essential for mathematics.

By integrating shape learning with STEM, preschoolers can develop a holistic understanding of how shapes play a role in various aspects of their world. This approach lays a strong foundation for future STEM education while making learning engaging and meaningful.

Helpful STEM Resources

Here are a few resources that will help you introduce STEM more effectively to your kiddos or students and feel confident yourself when presenting materials. You’ll find helpful free printables throughout.

  • Engineering Design Process Explained
  • What Is An Engineer?
  • Engineering Words
  • Questions for Reflection (get them talking about it!)
  • BEST STEM Books for Kids
  • 14 Engineering Books for Kids
  • Jr. Engineer Challenge Calendar (Free)
  • Must Have STEM Supplies List

Get Your Free Printable Shape Activity!

shape based activity recognition

25 Fun Shape Activities

Check out these easy hands-on ways for kids to learn about shapes.

Go On A Shape Hunt

Turn shape recognition into a treasure hunt by searching for objects of different shapes around the house or in nature. Kids can match found objects to shape cards.

shape based activity recognition

Sort Shapes

Provide a collection of toys or objects in various shapes and colors. Encourage children to sort them into corresponding categories based on the type of shape they are.

Trace Around Shapes

Give kids shape templates or stencils along with crayons or markers. They can trace the outlines to strengthen fine motor skills and reinforce shape knowledge.

Make Playdough Shapes

Mold playdough into different shapes using hands or shape cutters. This hands-on activity promotes sensory exploration and fine motor development.

shape based activity recognition

Create A Shape Collage

Gather magazines, colored paper, and glue. Kids can cut out shapes and arrange them into a creative collage.

Play With Shape Puzzles

Craft simple puzzles by cutting shapes out of cardboard and have kids match the pieces to the corresponding shapes on a board.

Homemade Shape Stamps

Transform everyday objects like sponges or bottle caps into stamps. Dip them in paint and create shape patterns on paper. See our art activity with shape stamps.

shape based activity recognition

Shadow Play

Utilize a flashlight to cast shadows of toys onto a surface. Challenge kids to identify the shapes based on the shadows. Check out more shadow activities for preschoolers.

Shape Baking

Incorporate shape learning into the kitchen by using cookie cutters to make shape-themed treats. Or cut fruit like watermelon into shapes for a yummy treat!

Nature Shapes

Venture outdoors to collect leaves, sticks, and stones. Use these natural items to create shapes on the ground, blending art with nature.

Outdoor Chalk Shapes

Head outside with sidewalk chalk ( here’s how you can make your own ) and draw various shapes on the ground. Preschoolers can hop from shape to shape, reinforcing shape recognition.

shape based activity recognition

Shape Books

Share picture books that feature shapes. Encourage kids to identify shapes within the illustrations. See our list of shape books above.

Shape Scavenger Hunt

Create a list of shapes for kids to find in their surroundings. They can check off each shape as they spot it.

Make Shape Puppets

Craft paper bag puppets with shape features like triangle noses or circular eyes. Kids can engage in imaginative play while reinforcing shape recognition.

Shape Play with Toys

Explore toys specifically designed for shape learning, such as shape sorters, puzzles, and magnetic shape sets.

Shape Building

Use building blocks, Lego, or other construction toys, or even straws and plasticine to build structures using different shapes. This fosters spatial understanding and engineering skills. Check our list of simple building activities.

shape based activity recognition

Collaborative Shape Art

Work together on a large piece of paper to create a collaborative art project using a variety of shapes.

Shape Memory Game

Play a memory game with cards featuring shapes. Kids flip the cards and match pairs of identical shapes.

Shape I Spy

Play “I-Spy” by describing shapes, prompting children to find objects that match the description within their surroundings.

Shape Hunt in Art

Explore famous artworks and identify shapes within the compositions. Discuss how artists use shapes to create images. For example; circles or rectangles .

shape based activity recognition

Play Shape Bingo

Develop a shape-themed bingo game. Instead of numbers, children mark off the shapes called out.

Shape Obstacle Course

Set up an obstacle course where each station features shape-related challenges. This promotes physical activity while reinforcing shape recognition.

Shape Songs and Rhymes

Sing songs or chants about shapes to make learning more interactive and memorable.

Shape Patterns

Create simple shape patterns using different colors and sizes. Extend the pattern and encourage kids to continue the sequence.

shape based activity recognition

Printable Shape Activities Pack

Who’s it for.

Perfect for 3-to 6-year-olds in a preschool classroom, learning center, or home. Easy to set up and use with one kiddo or turn into a center for the classroom. This themed pack includes a book and supply list to help you easily set up the 14 shapes activities.

⭐️ Tip:  Grab the  Preschool Themes Growing Bundle  and save money! ⭐️

What’s included.

  • Building Shapes
  • Seek and Find
  • Shape Printmaking
  • Sensory Sorting
  • What Shape is it?
  • Shapes Crown
  • Writing Center
  • Tracing Cards
  • Matching Game
  • Cover Up! Game

There are so many fun ways to explore shapes!

14 Comments

Great activity!

This is a terrific idea, I love the tactile nature of it. Fun and learning all rolled up together. Thank you!

Thanks a lot for this post. We are starting our homeschooling activities with my 8, 5, and 3 year sons. This is a great way of getting involved with geometry for the little ones.

Terrific! Glad you can use it!

Hi, I was wondering where do you print the shapes from?

This is where I found it:

“I printed off some geometric shape patterns from this site as it had some different sizes to offer” . I then cut them out and traced them onto the foam sheets. I made as many as I could so we would have lots of variety and colors! I set them out with popsicle sticks and magnetic numbers for an invitation to play and explore geometric shapes!”

  • Pingback: Hands On Geometry Learning Activities for Kids of All Ages
  • Pingback: 21 Quick STEM Activities for When You're In a Rush
  • Pingback: Math Activities Kids Love - FSPDT

This is an interesting activity that parents can teach their kids. Geometry and shapes, kids will like to spend time with such things. Even I shall try my hands on it. I really loved the new concept and will definitely share it with my friends too. I thank the author for this!

Thank you so much! It’s a great open ended math and STEAM activity for all ages!

I think this is a fantastic post. Geometry always looked boring in the books but now when you share a new way to learn, absorb and grasp geometry rules, it appears quite appealing to me and it may to the kids too. Tell me where to buy some such interesting games too.

  • Pingback: Learning about Shapes Math Activity | 123 Homeschool 4 Me
  • Pingback: 15 Fun Math Activities for Toddlers and Preschoolers 2019 | Entertain Your Toddler
  • Pingback: 15 FUN Math Activities for Kids! – The Homeschool Resource Room

Comments are closed.

~ Projects to Try Now! ~

shape based activity recognition

  • Math for Kids
  • Parenting Resources
  • ELA for Kids
  • Teaching Resources

SplashLearn Blog

10 Best Strategies for Solving Math Word Problems

5 Easy Tips & Tricks to Learn the 13 Time Table for Kids

How to Teach Number Sense to Kids: Step-by-Step Guide

How to Teach Decimals: A Step-by-Step Guide

How to Teach Fraction to Kids – 11 Best Activities

How to Choose Best School For Your Kid: 12 Best Tips

Why Kids Get Bored at School: 10 Tips to Keep Them Interested

11 Best Writing Apps for Kids

Homeschool vs Public School: 12 Tips on How to Choose One

15 Essential Life Skills Activities for Kids: Beyond ABCs

20 Animals That Start with “U”

70+ Easy Opposite Words for Kids in 2024

12 Animals that Start with K

12 Animals That Start With ‘E’: From Elephants to Eels

60 Best Essay Topics for Kids: Nurturing Young Minds

25 Best Websites for Teachers

10 best lesson planning apps for teachers.

15 Best Literacy Strategies for Teachers to Use in the Classroom

How to teach 4th Grade Kids: 25 Best Tricks & Tips

How to Teach Addition to Kids: From Counting to Calculating

15 Best Shape Activities For Preschoolers in 2024

Vector image of kids and shapes

1. Shape Hunt in Your Home

2. shape sorting game, 3. nature walk shapes, 4. playdough shape creations, 5. shape collage art, 6. shape puzzle time, 7. shadow shapes game, 8. shape bingo, 9. shape stamping, 10. shape obstacle course.

Do you ever look around and notice shapes all around you? You may see a round clock on the wall, a square picture frame, or even a triangle-shaped pizza slice! Shapes are everywhere, and guess what? They’re super important for kids to learn about when they’re little.

SplashLearn: Most Comprehensive Learning Program for PreK-5

Product logo

SplashLearn inspires lifelong curiosity with its game-based PreK-5 learning program loved by over 40 million children. With over 4,000 fun games and activities, it’s the perfect balance of learning and play for your little one.

So, let’s explore some of the best shape activities for preschoolers. In this blog, we will discuss some simple and fun ways to help preschoolers understand shapes like circles, squares, triangles, rectangles, and ovals. 

Shapes might seem like child’s play, but they’re a big deal. Learning about shapes isn’t just about knowing the names; it’s about sharpening young minds and setting the stage for learning all kinds of stuff in the future. So, let’s get started and make learning shapes a fun adventure for your preschooler!

Basic Introduction to Shapes

Have you ever played with your food, arranging your pizza toppings into a triangle or stacking your sandwich bread in a square? Shapes are like our secret code for understanding the world around us!

  • Circle: Think of round things, like a cookie or a bicycle wheel. They’re circles because they don’t have any sharp corners.
  • Square : Squares are like building blocks with four sides, all the same length. Picture a chocolate bar with four equal pieces – that’s a square!
  • Triangle: Triangles are those pointy shapes, like a slice of pizza. They have three sides and three corners, like a wizard’s hat.
  • Rectangle: Books are often shaped like rectangles – longer on one side, shorter on the other.
  • Oval: An oval is like a stretched-out circle, like an egg.

These shapes are all around us, from the sun in the sky to the windows in your house. Now, let’s have some simple and playful fun with them!

15 Best Hands-On Shape Activities for Preschoolers

Shapes are more than just simple figures children draw; they are foundational elements of mathematics that play a crucial role in a child’s cognitive development . According to the National Association for the Education of Young Children (NAEYC), preschool teachers can create an environment where children are eager to explore and learn about math, particularly shapes and space. The Common Core State Standards for Mathematics (CCSSM) emphasizes the importance of children understanding shapes by the time they leave kindergarten , highlighting activities such as analyzing, comparing, creating, and composing these shapes.

With this in mind, we’ve curated a list of engaging prek shape activities, each designed to nurture their curiosity and enhance their spatial understanding. Dive in and discover the world of shapes with your little ones!

A young kid holding two balls

Transform your home into an exciting shape treasure hunt for preschoolers, encouraging them to find shapes in everyday objects. This engaging activity sparks curiosity and helps children recognize and distinguish various shapes.

Materials Needed: None

Instructions: Ask your child questions like, “Can you spot a square door or a round plate?”

Kid sorting shapes

Make learning shapes hands-on and enjoyable with this interactive game. In the Shape Sorting Game, preschoolers sort everyday objects by their shapes, reinforcing their understanding of circles, squares, triangles, and more.

Materials Needed: Small objects (e.g., toy cars, buttons)

Instructions: Prompt your child to sort items, saying, “Let’s gather all the circles here and the squares there.”

Kid walking with father in park

Combine outdoor exploration with shape recognition during a nature walk, making learning shapes a delightful outdoor adventure. Preschool shapes activities like this encourage children to spot shapes in leaves, rocks, and clouds while connecting with the natural world.

Instructions: While walking, identify shapes in leaves, rocks, or clouds, saying, “Look, a triangle-shaped leaf!”

Kids making shapes from playdough

Source: @5minutesformom

Have a blast with playdough as your little one shapes basic preschool shapes like circles, squares, triangles, and rectangles. It’s a hands-on and colorful way to introduce and explore shapes while sparking creativity.

Materials Needed: Playdough in various colors

Instructions: Shape playdough into different forms and ask, “Can we make a rectangle or a triangle?”

A shape collage

Source: @pinterest

Shape Collage Art is a hands-on and artistic way for preschoolers to explore shapes found in the world around them. It’s a simple and enjoyable pre-k shapes activity that lets creativity shine. Create a beautiful collage using cutouts from magazines. 

Materials Needed: Old magazines, glue, paper

Instructions: Cut out shapes from magazines and arrange them to make a unique shape collage.

A handmade shape puzzle

Source: @proeves

Enjoy a fun shape activity for preschool as you create your own puzzles . Draw shapes on cardboard, cut them into pieces, and solve them together. It encourages problem-solving skills and helps preschoolers become familiar with basic shapes.

Materials Needed: Cardboard, markers, scissors

Instructions: Draw shapes on cardboard, cut them into puzzle pieces, and have your child assemble the puzzle.

Kid with parents using light to create shadow in wall

Source: @raisingchildern

The Shadow Shapes Game is a simple and one of the most fun pre-k shape activities that combines play and learning. Use a flashlight to create interesting shadows and guess the shapes they form. It’s a fun way to learn about shapes in the world around you.

Materials Needed: Flashlight

Instructions: Shine a flashlight on objects and have your child guess the shape of each shadow.

Shape bingo sheet

Source: @twinkl

Shape Bingo is an enjoyable and straightforward way to make geometry activities for preschoolers entertaining and educational. It turns learning shapes into a playful game that kids will love. Replace numbers with shapes on bingo cards and enjoy a fun-filled way to reinforce shape recognition.

Materials Needed: Bingo cards with shapes (you can make them)

Instructions: Play a simple bingo game, replacing numbers with shapes and enhancing shape recognition.

Shape stamp on paper using sponge

Source: @mothercould

Shape Stamping is an easy and enjoyable way for preschoolers to explore shapes and colors through art. It encourages creativity and helps them become familiar with various shapes.

Materials Needed: Sponges, paint, paper

Instructions: Dip sponges in paint and stamp them onto paper to make colorful shape art.

Kids playing shape obstacle course

Source: @rehabshop

The Shape Obstacle Course combines physical play and shape activities for preschoolers, making learning about shapes an active and enjoyable experience. It helps children associate shapes with movement and promotes gross motor skills .

Materials Needed: Tape, cushions

Instructions: Create a tape outline of shapes on the floor, and encourage your child to jump or crawl through them.

11. Shape Treasure Map Adventure

Kid looking at a treasure hunt map

The Shape Treasure Map Adventure adds an element of excitement and exploration to shape activities for preschoolers while making learning shapes feel like a thrilling treasure hunt. It’s a creative and engaging way to reinforce shape recognition skills in a playful manner.

Materials Needed: Paper, markers or crayons, small objects (as treasures)

Instructions: Draw a simple map with shape landmarks and have your child follow the map to find hidden treasures matching the shapes.

12. Shape Storytime

Mother reading to daughter

Shape Storytime is a simple and enjoyable way to introduce shape activities for preschoolers through the magic of storytelling . It helps children relate to shapes in familiar and engaging contexts while fostering a love for reading .

Materials Needed: Picture books with shapes

Instructions: Read a story with shape-themed books, pointing out and discussing the shapes you encounter.

13. Build with Blocks

Kid playing with building blocks

Build with Blocks is a hands-on, interactive, and simple way to engage in shape activities for preschoolers. It fosters creativity, shape recognition, and bonding through constructive play.

Materials Needed: Building blocks

Instructions: Build simple structures with blocks, discussing the shapes used in the process.

14. Shape Balloon Fun

kid playing balloon catch with parents

Shape Balloon Fun is a light-hearted and easy way to introduce shapes during playtime. It adds a fun twist to learning about shapes and encourages simple interactions that help preschoolers grasp the concept.

Materials Needed: Balloons

Instructions: Inflate balloons and talk about their shapes while playing catch.

15. Shape Cooking Adventure

kid cooking with mother

Combine shapes and delicious treats in a Shape Cooking Adventure that’s both fun and tasty. It’s a hands-on activity that brings the joy of baking into shape activities for kids.

Materials Needed: Cookie dough, cookie cutters

Instructions: Bake shape cookies together using various cookie cutters, and discuss the shapes as you enjoy your homemade treats.

These preschool shape activities and crafts are designed to be simple, engaging, and perfect for introducing your little one to the wonderful world of shapes. 

4 Extension Ideas and Variations To Teach Your Preschooler Shapes

Now that we’ve explored some exciting preschool shape activities, it’s time to take the fun and learning even further with some creative extension ideas and variations. These simple tweaks and additional concepts will add depth to your shape activities and keep your child engaged and eager to learn. Let’s delve into these innovative extensions:

1. Incorporating Colors with Shapes

Enhance the learning experience by incorporating colors into your shape activities. You can ask questions like, “Can you find a red circle or a blue square?” This combination of shapes and colors adds an extra layer of exploration.

2. Using Real-Life Objects for Shape Recognition

Take shape recognition to the next level by integrating real-life objects. When you’re out and about, point out things that match the shapes your child is learning. For example, “Look, that stop sign is an octagon!”

3. Combining Shapes to Create New Objects

Encourage creativity and critical thinking by combining basic shapes to create new objects or creatures. Challenge your child to make a house using squares and triangles or an animal using circles and ovals.

4. Incorporating Technology with Digital Shape Activities

Embrace technology by exploring digital shape activities and apps designed for preschoolers. Many interactive apps offer shape-recognition games and puzzles, making learning more engaging.

You can customize the learning experience to suit your child’s interests and needs by infusing these extension ideas and variations into your shape activities. These adaptations add depth, creativity, and excitement to the world of shapes, ensuring that your preschooler continues to explore and grow in their understanding of this fundamental concept.

Shapes and spatial reasoning are foundational blocks in early childhood education. Dr. Sue Gifford’s insights from NRICH emphasize the profound connection between a child’s ability to understand and manipulate shapes and their mathematical development. Activities that challenge children to analyze, combine, and rotate shapes, such as puzzles and construction play, are fun and pivotal in enhancing their spatial thinking. The language we use, whether discussing the ‘curviness’ of a circle or the ‘ edges ‘ of a square, significantly deepens their understanding of shapes. Furthermore, physical activities that involve navigating spaces, constructing routes, or even simple tasks like stacking help children relate to shapes in the world around them. 

Placing shapes at its heart is essential as we craft our early years curriculum. By doing so, we’re not just teaching children about geometry but equipping them with a comprehensive toolkit of cognitive skills that will serve them well in all areas of learning.

Frequently Asked Questions (FAQs)

How do you teach shapes in preschool.

Teach shapes in preschool through hands-on activities like shape sorting and creative playdough shaping.

What are the shape concepts for preschoolers?

Shape concepts for preschoolers include recognizing common shapes like circles, squares, triangles, and rectangles, as well as understanding their properties.

How do you make shapes interesting?

Make shapes interesting for preschoolers by incorporating games , art, and outdoor exploration to engage their curiosity and creativity.

shape based activity recognition

15 Best Empathy Activities for Kids to Foster Kindness

13 Best Black History Month Activities for Kids

24 Best Sensory Activities for Preschoolers

Preschool

Most Popular

A working mom and her daughter in the bedroom, Mom is working while daughter is playing with her toys.

101 Best Riddles for Kids (With Explanation)

shape based activity recognition

15 Best Report Card Comments Samples

Good vibes quotes by SplashLearn

40 Best Good Vibes Quotes to Brighten Your Day

Recent posts.

Online education depiction

Math & ELA | PreK To Grade 5

Kids see fun., you see real learning outcomes..

Watch your kids fall in love with math & reading through our scientifically designed curriculum.

Parents, try for free Teachers, use for free

Banner Image

  • Games for Kids
  • Worksheets for Kids
  • Math Worksheets
  • ELA Worksheets
  • Math Vocabulary
  • Number Games
  • Addition Games
  • Subtraction Games
  • Multiplication Games
  • Division Games
  • Addition Worksheets
  • Subtraction Worksheets
  • Multiplication Worksheets
  • Division Worksheets
  • Times Tables Worksheets
  • Reading Games
  • Writing Games
  • Phonics Games
  • Sight Words Games
  • Letter Tracing Games
  • Reading Worksheets
  • Writing Worksheets
  • Phonics Worksheets
  • Sight Words Worksheets
  • Letter Tracing Worksheets
  • Prime Number
  • Order of Operations
  • Long multiplication
  • Place value
  • Parallelogram
  • SplashLearn Success Stories
  • SplashLearn Apps
  • [email protected]

© Copyright - SplashLearn

  • Skip to main content
  • Skip to footer

Miss Kindergarten

Kindergarten Teacher Blog

Hands-On Shape Activities for Preschool

May 29, 2022 misskindergarten Leave a Comment

Shape identification is one of my favorite concepts to teach preschoolers because are so many hands-on and engaging activities!  No matter the age of your young learners, there are multiple ways that they can explore and learn about basic shapes. Read on for some of my favorite hands-on shape activities for preschool!

Shape activities for preschool - Car shape tracing activity

Shape Matching Activities for Preschool

One of the first steps that I like to take in shape recognition is having students use their visual discrimination skills to match shapes.  It’s helpful for students to explore the differences and similarities in shapes as they identify matches. Here are some of my favorite ways to help students match shapes and explore shape orientation.

  • Task Cards:   For example, students can match the shaped wheels to the correct bus
  • Shape Boards and Puzzles: Explore shape matching and orientation by completing puzzles
  • Pattern Blocks : Students can match pattern block shapes based on pattern block cards
  • Concentration: Play a shape matching game, adjusting difficulty by adding or removing cards
  • Partner Up: Give each student a shape card, then have them find the classmate with a matching shape

Shape Building Activities for Preschoolers

Exploring shapes by building them is another great opportunity for preschoolers to learn about shapes.  A class of preschoolers often displays a wide variety of fine motor skills. This is why shape building activities are perfect for preschool! They provide additional fine motor practice before moving on to tracing and writing shapes. 

Two shape building cards are being used with play dough

  • Stick Shapes: Use popsicle sticks or toothpicks to build shapes
  • Play Dough Mats: Roll out play dough to create a variety of shapes with these free mats
  • Pattern Blocks: Build a larger shape using pattern blocks
  • Shape Building Mats: Use any manipulative you have on hand to create shapes on these mats
  • Geoboards: Explore shapes using rubberbands and geoboards 
  • Flexible Shapes: Build shapes with flexible materials like pipe cleaners or Wiki Sticks

Activities to Identify Real-Life Shapes

As students become more familiar with shapes, it’s important for them to see that there are shapes all around them!  Here are some fun activities for students to see shapes in real-life objects.

A shape center is in use on a desk, with real-life objects being sorted into groups based on their shape.

  • Shape Matching Cards: Match a picture of a real-life object with the corresponding shape
  • Play “I Spy” : Have students take turns thinking of a shaped object in the room while everyone else guesses
  • Show and Tell: Have a shape of the week to encourage students to see shapes in the world around them
  • Shape Hunt: Go on a shape hunt in nature or around the school grounds
  • Sensory Table Sort: Add buttons or other shaped objects to a sensory table for sorting

Shape Tracing and Writing Activities

We often think of paper and pencil when it comes to tracing and writing. However, there are many other activities that will help our young students practice proper shape formation before they ever need to pick up a pencil!  These activities are especially helpful for young students who are still developing their pencil grasp.

students love learning shapes with these car themed shape cards

  • Shape Roads : Shape road task cards help students explore the concept of tracing along a path. No cards? Use painter’s tape to create shape roads on the ground.
  • Water Painting: Draw shapes with sidewalk chalk, then have students water paint over them
  • Sensory Bag: Draw a shape on the outside of a sensory bag, then have students trace the shape with their finger
  • Paint and Outline: Stamp paint using shaped blocks or sponges, then have students outline with a marker once dry
  • Dry Erase Task Cards: Laminate shape task cards for tracing with a dry erase marker
  • Highlighting Shapes: Trace shapes with a highlighter
  • Cookie Cutters: Use cookie cutters as stencils or as a stamp and trace activity

Activities to Practice Shape Names and Attributes

As students explore shapes, they are naturally introduced to the shape names and will start to have a basic understanding of the shape attributes.  Depending on the age and needs of your students, they might be ready for more explicit instruction about shape names and attributes.  These fun activities will help them be kindergarten ready!

Three Shape Bingo cards with the corresponding cards for playing the game.

  • Play Shape Bingo: Call out a shape name (or describe the attributes only) and have students find the shape on their bingo cards.
  • Shape Hunt : Attach paper shapes to the bottom of a small box or activity tub, cover with salt, then have students use a paintbrush to hunt for shapes, naming each shape as they find it.
  • Mystery Shapes : Draw a variety of shapes on white paper using white crayon.  Have students paint their paper using watercolor paint, then name the shapes they see.
  • Shape Mini Books: Have students practice writing shape names and describing attributes with printable shape books.  They love to take these home to share with their families!
  • Shape Poems: Use these shape poems to help students remember shape names and attributes.  These catchy rhymes are always a hit in preschool classrooms!

Shape Activities for Preschool

Are you looking for hands-on shape activities that are ready to print and use in your preschool classroom?  I have some resources that you’ll love!

The first is a bundle of morning work tub activities for preschool .  This resource is full of engaging activities that are perfect for morning work!  Plus, these activities are grouped as beginning, middle, and end of year tasks that grow with your students.  In this bundle, you’ll find a variety of shape practice activities for your students, as well as literacy and math activities for preschool! You can find this resource in my shop .

Morning Tubs Preschool Bundle

Are your students ready for more of a challenge as they approach kindergarten age?  If so, you’ll love this 2D Shape Centers and Activities bundle for young learners.  These hands-on centers are perfect for young learners who are ready for more of a challenge as they prepare for kindergarten.  Students will be able to practice identifying, building, and writing 2D shapes in a variety of ways. This resource is also available in my shop .

Save These Shape Activities for Preschool

If you’d like to refer back to this list of shape activity ideas for preschool, be sure to save this post!  Just add this pin to your favorite preschool board on Pinterest so you can quickly find this post whenever you need it.

Shape Activities for Preschool

Number Formation Poems

You may also enjoy these posts.

Slide1.jpg

Check Out These Resources

shape based activity recognition

March Activity Bundle

shape based activity recognition

Addition and Subtraction Practice Activities BUNDLE

shape based activity recognition

Phonics Activities CVC Words MEGA BUNDLE!

St. Patrick's Day Math and Literacy Centers

St. Patricks Day Centers for Kindergarten

Shop All Resources

shape based activity recognition

Hello, I’m Hadar

Welcome to Miss Kindergarten. I’m so happy you’re here!

If you are looking for hands-on, engaging kindergarten activities, you came to the right place! I’m here to save you time by sharing tried and true kindergarten resources, and hopefully spark some ideas for your own kindergarten lesson plans!

Whether you need ideas to teach reading, sight words, math, or even some fun crafts, I have you covered. My ultimate goal is to help passionate educators and parents to young kids gain their valuable time back!

If you want to stay connected with Miss Kindergarten, please follow me on social media and be sure to sign up for the newsletter below.

More About Me Contact Me

shape based activity recognition

Follow Miss Kindergarten on Instagram!

Fun ideas at your fingertips! Get daily inspiration, tips, and ideas, plus special sales and freebies! Connect with Miss Kindergarten on Instagram.

Helpful Links

  • Free Resources

Popular Posts

Teaching Long Vowel Sounds with Decodable Passages

Teaching Long Vowel Sounds with Decodable Passages

Teaching Long Vowel Sounds with Decodable Passages

Winter Holiday Activities for Kindergarten

Teaching Long Vowel Sounds with Decodable Passages

2D Shape Poems and Rhymes

New in the shop.

March Activity Bundle

March Low Prep Kindergarten Centers Math and Literacy Centers

March Activity Bundle

March Morning Work for Kindergarten

shape based activity recognition

  • Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar

Fantastic Fun & Learning

Fun learning activities and things to do with kids

Shape Activities

By Shaunna Evans 2 Comments · This content may contain affiliate links.

There are so many ways to incorporate  hands-on math activities for kids into playful fun and learning. In this collection we’re specifically looking at creative shape activities to help kids learn about 2D shapes.

Shape Activities for Kids

Make  shape rocks  for kids to use for play and learning. Fun-A-Day shares some excellent ideas for using them.

Scavenger hunts are always a hit around here. Buggy and Buddy has a free printable to take along with you for a shape scavenger hunt at the playground.

Making shape collages like these from Homegrown Friends can be a great way to expose young kids to shapes. The process also promotes creativity.

Make your own geoboard with this tutorial from Crayon Box Chronicles. I love the use of fabric loops in this one!

2D Shape Activities for Kids

Toddlers and preschoolers will love the added sensory element of water and bubbles in this scooping shapes sensory bin .

Little creators can turn 2D shapes into a bird or another favorite creation in this 2-D shape art invitation .

Use play dough and these free printable shape mats to make shape pet “food” for pet theme math fun!

Turn shapes into collage art monsters in this creative monster art shape activity .

2D Shape Puzzles with real photographs

Help children begin to recognize 2D Shapes in real objects with these free printable photo shape puzzles .

Shape Activities 2

Fun-A-Day has another homemade geoboard that is super easy to make and great for letting kids create their own shapes.

This Foam Lacing Shapes activity is great fine motor practice and makes a great busy bag.

This Build a Bracelet activity is great for shape recognition and fine tuning your child’s fine motor skills.

This inexpensive Cardboard Shapes activity easy toddler craft is a cheap way to teach your toddlers shapes.

shape based activity recognition

Learn to recognize shapes in everyday objects with these quick prep free printable 2D shape puzzles .

Shape Activities 3

This fun Trace the shapes art project is a great fine motor activity and a great way to do some art while learning shapes.

This simple Learn about shapes using paper plates and yarn activity is so easy to set up and will be so much fun for your little ones.

These Shape Play Dough Mats will teach your kids the basic shapes so they can start to recognize and learn them.

If you are looking to get your kids active and moving, you can’t go wrong with this Shape Hopscotch activity.

RELATED: 25 Shape Books for Kids

More math and science activities with fizz, pop, bang.

Fizz, Pop, Bang! Playful Science and Math Activities  is designed to bring hands-on fun to math and science play. It’s full of engaging and powerful learning opportunities in math and science, shared through ideas that incorporate art, play, sensory learning and discovery, for a whole-brain approach.

It includes 40 educational projects and 20 printables including a set of build-your-own 3D shape blocks, engineering challenge cards and a range of math games.

Learn more about  Fizz, Pop, Bang!  or you can  buy it now !

Buy Now-Fizz Pop Bang ebook with 40 projects for ages 3 to 8 and 20 free printable resources

Reader Interactions

Mary Catherine

February 19, 2014 at 6:24 pm

Thanks so much for including my rainbow rocks! I love seeing the other shape ideas you have here — can’t wait to check out the rest!

February 19, 2014 at 2:32 pm

Thank you for featuring our DIY Geoboard with fabric loops. So many great activities. Pinned.

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Early Impact Learning

12 Fun Preschool Shapes Activities to Try

Exploring and identifying shapes is an important learning experience for preschoolers. Shapes are all around us and understanding them makes children make sense of the world. By learning shapes, kids develop their memory, visual recognition, spatial awareness and problem-solving. They should learn to identify basic shapes and recognize them in everyday objects from an early age.

shape based activity recognition

Table of Contents

Creative Shape Activities for Preschool

There are many shape activities for preschool that you could engage in with your children to enhance their creativity and imagination. Young kids can explore shapes by building with blocks, sorting objects and playing shape-related games. Teaching shapes is a great way to learn new vocabulary and develop early math skills while having fun. Below are three of my favorite shape activities that I have used in my classroom. I love to watch as the children’s understanding grows while setting a strong foundation for future learning.

shape based activity recognition

1. Play Dough Shape Activities for Preschool

Prepare different colored play dough and some pictures or real shapes. Go through their characteristics by introducing words like corners and edges. Let the children explore shapes and play with them for a few minutes.

Then, ask the children to warm up the play dough and open and close their hands. Teach shapes by showing them how to make different shapes using the play dough. Can they make the basic shapes of a circle, square or triangle?

shape based activity recognition

2. Shape Collage Creation

For this activity, you will need:.

An A3 sheet for each child, colored cards, scissors and gluesticks.

Begin by introducing the concept of collage to the children. Explain that a collage is a piece of art created by putting different materials together, even if they overlap. Show them an example you created, and go through all the shapes you have already cut out.

If the children are older, they can draw and cut the shapes themselves.

Can they pick out a square? Can they find a yellow triangle? Go through all the names of the shapes you will work with.

After all the shapes have been cut out, model how to stick the shapes on the paper to create a collage.

Allow children to display their collages and talk about them.

shape based activity recognition

3. Shape Printing

You will need:.

Sponges or foam cut in different shapes, poster paint and sheets of paper or card.

shape based activity recognition

Shape printing is a very fun way to teach shapes! First, explain to the children that they will be doing a fun geometry activity where they will be using shapes as stamps to create unique patterns and designs.

Engage them in a discussion about the shapes made from sponges or foams and see if they can recognize them. Change the colored paint you will be working with, too. Can children name all the colors? Do they know which colors can be mixed to obtain a different color?

Model how to dip the sponges in the paint and ensure they are evenly coated. Encourage children to experiment with different combinations of shapes, colors, textures and arrangements to express their creativity.

Can they create patterns?

shape based activity recognition

After the activity, open a class discussion on their creations and ask them about their favorite prints.

You can even create a collaborative shape print mural by sticking all the pieces of paper together.

4. Shape Block Towers

shape based activity recognition

For this shape activity you will need:

Building blocks in different shapes, sizes, and colors. Cuisenaire rods work fine, too.

shape based activity recognition

For older children, include shapes such as cubes, rectangular prisms, cylinders, and pyramids. Begin by explaining that they will be using different shapes of blocks to build towers.

Engage the students in a discussion about the different shapes of blocks. Show examples and let your little learners identify shapes.

Then model how to build a tower by using the shapes at hand.

Mix things up. Challenge some children by allowing them to use a certain number of blocks, include specific shapes, or build the tallest tower they can.

Can children use different combinations of shapes, create patterns, or alternate colors? Encourage discussion.

Once the towers are built, talk about the children’s creations. Ask questions, such as which shapes were the easiest or most challenging to work with, and what they learned from the activity.

As a plenary, ask children to describe their towers in the classroom. Expand on their imagination by asking them questions such as: ‘Who lives in your tower?’, ‘How many rooms does it have?’, ‘What is there around your tower?’

Problem-Solving with Shapes

Problem-solving with shapes in the classroom offers students an opportunity to develop critical thinking skills while learning shapes and exploring geometric concepts.

Here are some of my favorite shape activities I have used in my classroom:

5. Shape Puzzle Challenge

You will need different shape puzzles. Explain that they will have to fit different shapes in the correct gaps.

Gather age-appropriate puzzle pieces in various shapes and different puzzle boards. Discuss each shape, and talk about how they can fit together with another shape, or which gap they correspond to.

A tangram is a great resource to use, like in the picture below:

shape based activity recognition

For older children, introduce pentagons, hexagons and octagons.

Show the children how to fit the pieces together on the puzzle board. Distribute the puzzle pieces to the children or allow them to take turns selecting pieces. Encourage them to work together, guiding them as needed to fit the pieces together correctly.

Let the children explore the puzzles individually or in pairs. Observe and provide support as necessary.

As a plenary, ask children which shapes were the easiest or most challenging to piece together. Why?

This shape activity is great to develop fine motor skills .

shape based activity recognition

6. Shape Memory Game

Learning shapes can be enjoyable and engaging with the right activity, such as this fun shapes game.

Cards that have different shapes printed on them. You can make the cards yourself, or find them online and print them.

How to play this shape activity:

Start by placing the cards face down in a grid formation. Start with only 4 cards first, especially if the children are very young, then gradually increase to six or even eight.

The first player flips over two cards, trying to find a matching pair of shapes. If successful, they keep the cards and take another turn. If not, they flip the cards back over, and it becomes the next player’s turn.

The game continues until all the matches are found. The player with the most pairs at the end wins. This game enhances memory skills , shape recognition, concentration, and fine motor skills.

It can be adapted for different levels by adjusting the number of cards or introducing more complex shapes.

7. Shape Sorting Bonanza

This is one of the most fun shape activities for preschool that I recommend you try with your young kids or students.

Different shapes that are made from different materials eg. wood, plastic or card.

To play a shape-sorting bonanza game with preschool kids, gather a variety of objects or flashcards representing different shapes.

shape based activity recognition

Explain that they will have to sort the shapes according to different criteria, for example by color, material, shape, roughness, thickness, etc.

Make a list of all the different sorting criteria they could use. Then model the first example yourself.

After, spread the shapes on a table or play area. Guide the children to sort the objects into corresponding shape categories, placing circles with circles, squares with squares, and so on.

As a challenge and a creative twist for more able or older children, can they find shapes in their environment such as a round clock or a square pencil/case or a rectangular sharpener or eraser?

Ask them to gather all the shapes they can see around them and sort them into different categories.

8. Shape Jumping

Masking tape and a clear portion of the floor.

shape based activity recognition

Make big shapes on the floor using masking tape. The children will be asked to jump from one shape to another one by one. Ensure you are teaching shapes by going through each one and making sure the children know the names.

Then ask the children to make a line and tell them which shape to jump on until they get to the other side. Alternatively, ask them to walk around the sides of a particular shape (on the masking tape) and ask them to count how many sides it has. Can they count the corners?

Learning Through Play

Children always learn best through play, and exploring shapes is no exception. When children manipulate shape puzzles, build with blocks, or create shape collages, they develop spatial awareness, problem-solving skills, and fine motor control. Playful shape activities encourage them to explore the attributes of shapes, compare and classify them, and understand their real-world applications. So let them learn shapes through play! Try these shape activities for preschoolers below:

9. Shape Sensory Bin

Sensory bins are amazing for preschool children . They usually can’t get enough!

shape based activity recognition

Prepare a sensory bin by filling it with a base material like colored rice, sensory beads or pasta.

Add a variety of objects and materials that represent different shapes, such as foam or wooden shapes, plastic shape molds, or shape-themed toys. Introduce the shape sensory bin to the children, explaining that they will be exploring shapes through sensory play.

Sensory Exploration

Encourage the children to use their hands to feel and manipulate the shapes in the bin.

Let them sort the shapes, identify and name them, and describe their characteristics (e.g. sides, corners).

Shape Search

Hide specific shapes within the sensory bin.

Guide the children to search for and identify the hidden shapes.

Encourage imaginative play by letting the children create scenes or stories using the shape objects in the sensory bin.

shape based activity recognition

Teach children the habit of tidying up the shape objects and returning them to the sensory bin.

10. Shape Scavenger Hunt

Begin by explaining the concept of a shape scavenger hunt to the children. Give them each a list of shapes to find and define the boundaries, whether it’s the walls of the classroom, the library or the garden. Explain that when they find a shape on their list, they have to bring it back to you, or collect it in a container.

Divide the children into small teams or pairs, ensuring each group has a list of shapes to find. Encourage the children to search the designated area for objects that match the shapes on their list.

shape based activity recognition

Have each team share their findings and discuss the shapes they found. Celebrate their accomplishments and reinforce their shape recognition skills.

As an extra treat, offer small prizes or certificates to encourage engagement and participation.

11. Shape Detectives (I-spy)

Children have to spot the shapes which are mixed in with many different other objects.

Tell the children they do not move around the room. Put many small toys and objects in front of them. Tell them that they have to spot the shapes and take turns to pick them out of the pile of toys. They have to name the shape, pick it up and say its color and properties.

shape based activity recognition

This can be a timed activity, as children have to give their maximum focus for a limited time period.

12. Shape Hunt in Nature

Shape Hunt is similar to the scavenger hunt, but the twist is the location. Ask children to find the shapes you have previously hidden in a forest, park, or school playground.

Make the boundaries very clear, and ask the children to collect the shapes they find.

Come together at the end to talk about the shapes they have found.

shape based activity recognition

In conclusion, engaging preschool children in shapes activities offers a multitude of benefits. These activities promote shape recognition, spatial awareness, fine motor skills , creativity, and cognitive development. By incorporating playful and hands-on experiences, we can nurture their curiosity, ignite their imagination, and lay a strong foundation for their ongoing learning journey.

WonderBaby.org

Helping Your Baby Reach Greater Wonders

5 Shape Recognition Activities to Boost Development

Natasha Combrink

  • Children start showing an interest in shapes around 2 or 3 years old, but you can teach your little one about them from birth through everyday interactions.
  • Your child can improve their visual-spatial abilities and spatial awareness by recognizing shapes. 
  • Shapes are a key component of the foundation kids need in math and science.  
  • There are several ways to help your child learn shapes. Some examples are pointing out shapes, using technology, and playing shape-recognition games. 

I started teaching our daughter about different shapes before she went to preschool. When her first report card came home, I could see the games we played and the songs we sang worked! Now, I’m doing the same with our son. Teaching kids about shapes in a fun and interactive way helps them retain what they learn. 

Play is learning, and there’s no better way to teach your kids about shapes. Once you understand the importance of shape recognition, you can find out how to play your way through all the shapes!

Shape Recognition: Why Kids Need to Learn Shapes

Recognizing shapes is an essential skill for children. It forms the foundation for many other areas of learning, like math, reading, and science. Here are some reasons why kids need to learn shapes:

Visual-Spatial Abilities: The ability to recognize and manipulate shapes is closely tied to visual-spatial skills. These abilities are important for reading maps, building blocks, and using technology. 

Spatial Awareness: Understanding shapes and their properties will help your child develop a sense of how objects relate to one another in specific spaces. This awareness will help them visualize and manipulate objects in their head. 

Problem-Solving: Recognizing and identifying shapes will help your little one approach problem-solving and reasoning in a structured manner. 

Enhancing Creativity: Recognizing and using shapes can also help your children’s creativity. They can use shapes to create pictures, designs, and other works of art.

Mathematical Skills: Shapes are a crucial component of geometry. Learning about shapes will lay the foundation your kids need for future math skills. 

Children kids play with educational toys, arranging and sorting colors and shapes.

When Is the Best Time to Teach Your Child About Shapes?

Children typically show an interest in shapes around 2 or 3 years old. This is a great time to teach them more about it. Preschoolers should identify and recognize basic shapes.

You should, however, teach shape recognition before this age. Children can start learning shapes in a fun and playful way from birth! You can point out shapes in their environment and identify them in their toys. 

You can also use everyday objects to show your child how different shapes look. Talking to your kids about shapes will help develop their awareness of them. 

You don’t need worksheets and complicated activities to help your children learn shapes. Dr. Kate Green (Ph.D. in Child Development) says that learning through play positively influences literacy and mathematic development in children. 

At What Age Should a Child Be Able to Identify Basic Shapes?

Every child develops at their own pace. Don’t be concerned if your kid takes a little longer to grasp the concept of shapes. Most children will identify basic shapes like circles, squares, and triangles by age 3 or 4.  

Your child will likely hit this milestone on time if you playfully introduce shapes. Kids should be interested and motivated to learn about shapes if you want them to retain the information. Hands-on activities and games will capture their attention and help them enjoy getting to know different forms . 

How to Help Your Child Learn Shapes

Preschool teachers often focus on shapes before teaching kids the alphabet. This is because forms build the knowledge children need to learn letters and how to read. You can help your child learn shapes long before they go to school. Below are some fun ways to do this!  

Help Your Child Learn Shapes by Using Real-Life Examples

The easiest way to teach shape recognition is by pointing out shapes in your child’s surroundings. This will help your child make real-world connections and reinforce their understanding of different shapes.  

You can point out square windows, round wheels, or the triangular shape of a pizza slice. Ask your little one to look for and identify different shapes as you do your daily activities. You can make a game out of it or challenge them by asking for a specific shape. 

If you have more than one child, motivating them with a prize is also a good idea!

Help Your Child Learn Shapes by Playing With Them

Playing with shapes is a fun and interactive way to learn. Give your child different types of shapes to play with. You can encourage older children to build structures with blocks and identify the shapes they use. Shape sorting toys and toddler puzzles are great for toddlers starting to learn shape recognition. 

If you have play dough, press out a few shapes with your cookie cutters and ask your kids to match them. You can also draw, color and paint different shapes. 

Help Your Child Learn Shapes by Using Technology

There are many educational apps and games available that help children learn shapes. Look for age-appropriate content that is fun and interactive. 

You can find games that teach shape recognition, sorting, and matching. Some games develop problem-solving skills and teach children how to draw shapes. For younger children, choose games with common shapes like the circle, the square, the triangle, and the rectangle. Preschoolers should play games introducing the oval, star, diamond, and heart. 

Note: Read feedback from other parents and limit screen time.  

Kid girl plays with educational toy indoor.

5 Shape Recognition Activities Kids Will Enjoy

You can do many activities with your children at home or on the go to teach them shapes. Below is a list of games you can play that will teach shape recognition while encouraging creativity, problem-solving, and practicing gross, fine, and visual motor skills . 

Shape Scavenger Hunt

You can create a shape-inspired scavenger hunt where your child has to find and identify shapes in your house or at the park. This activity will help little ones make real-world connections, practice problem-solving skills, and reinforce shape recognition. 

Here’s a simple guideline for this activity:

  • Make a list of all the shapes your children need to find. You can also give them shape cards that they have to match to real-world objects.
  • Set a timer to add a sense of urgency and get your kids excited!
  • Reward the winner with a small prize. If you only have one child, motivate them with their favorite treat or another activity once they complete the scavenger hunt. 

Shape Sticker Designs

Playing with stickers is a fantastic fine motor activity that’s easy to set up. Children can come up with great ideas when doing this task! 

Start by giving them stickers made from different shapes and pieces of paper. You can ask your child to create an image using stickers to represent different objects. They can use a circle for the sun or triangles for trees. Let them be creative and explain to you how their imaginary picture was made when they’re done. 

This activity encourages creativity and helps kids identify how shapes can be used to represent different objects. 

Shape Bingo

If you want to teach shapes to preschoolers, playing shape bingo is a fun way to do this. This game will help children recognize shapes, practice matching, and build their visual and fine motor skills .  

To prepare for this activity, create a worksheet with shapes or objects representing shapes. Use colors and pictures that will grab your child’s attention. 

Give your child the bingo sheet and ask them to place a sticker or other marker on the shapes when you call them out. You can show your kid pictures of shapes, say their names, or point to real-world objects.  

Sing Shape Songs

Singing shape songs is a lively and entertaining way to help children learn about and recognize shapes. This type of activity is especially beneficial for younger kids. Songs help kids engage their imagination and improve their vocabulary. 

Singing shape songs is one of the best ways to teach shapes. Your child will develop memory skills when they repeat information, like the names of shapes. Shape songs also expand vocabulary by introducing new words and concepts.  

Many shape songs involve movements and gestures. Making this fun will help your child stay engaged and excited to learn the words. 

You can choose between great shape songs like:

  • Shape Are All Around by Pinkfong Songs
  • The Shapes Song by Oh My Genius
  • Nashville Shapes Hoedown by Busy Beavers
  • Shape Song by CoComelon
  • Sing Along Shape Song by Bounce Patrol
  • Shape Song by Gracie’s Corner

Shape Sorting

Shape sorting games provide a fun way to practice shape recognition and fine motor skills. It also contributes to cognitive development and improves problem-solving abilities. You can get creative when setting up a shape-sorting game. Here’s an easy-to-create game that you can adjust to your kid’s age:

  • Give your child a basket containing different-shaped objects like wooden blocks , puzzle pieces, or cut-outs. 
  • Ask them to sort the shapes into categories like circles, squares, and triangles.
  • Make the game more challenging by asking them to color-sort the shapes too! 

This activity will practice categorizing and grouping while reinforcing their understanding of different shapes. 

The Secret to Learning Shapes: Play!

You don’t even have to look closely to figure out that rectangles, circles, squares, and triangles are around us all day. Kids who know the difference between these shapes can identify and organize visual information. This is one of the biggest reasons to start learning shapes early on. 

If you keep the activities you do around shape learning playful, I’m sure your kids will stay engaged and retain information. Remember to choose age-appropriate games and adjust your expectations for your child’s abilities.

Shape Recognition Activities to Boost Development

Related Posts

Close-up shot of young caregiver trying to comfort and calm down a crying child.

Development

How Fearful Avoidant Attachment Develops in Childhood

Is your child struggling with emotional regulation, incongruent behaviors, and boundaries with strangers? Fearful avoidant attachment may be to blame.

Mother obsessed with control practicing helicopter parenting style.

Development, Parenting

Distal and Proximal Parenting: Understanding the Difference

Understanding the history, differences, and strengths of proximal and distal parenting will help you decide what parenting approaches work best for your family.

Newborn baby with open mouth is in arms of his mother and looks up.

Autism, Development

My Baby’s Mouth Is Always Open. Is It an Autism Indicator?

If your baby’s mouth is always open, you might be concerned about a link to autism. However, there are many other reasons for a babies mouth to be open.

PKP-Logo-Recolor-2022

Ready to Make Circle Time Amazing?

Sign up for our FREE newsletter and receive my ebook 7 Circle Time Mistakes

Thanks for subscribing! Please check your email for further instructions.

Shapes Activities for Preschoolers

shape activities for preschool

One of the first math concepts that preschoolers learn is identifying shapes . They begin to distinguish among the different shapes and categorize items according to shape. They learn the names of shapes and their characteristics. They find shapes in everyday items. This collection of shape activities for preschoolers can lead preschoolers to explore shapes in all kinds of ways.

Learning Activities for Shapes in Preschool

These activities will help your preschoolers learn their shapes. These shape activities for toddlers, will work in your preschool classroom and your kindergarten classroom.

1. Road Shapes with Cars (Pre-K Pages) – 22 printable road shape mats to help your litte learners identify shapes.

2. Making Shapes with Play Dough (Pre-K Pages) – A fun, hands-on playful learning experience that uses play dough to teach shapes!

3. Pattern Block Shapes (Pre-K Pages) – Pattern blocks can actually help your little learners build a strong foundation for learning geometry later.

4. Making Shapes with Geoboards (Pre-K Pages) – If you haven’t tried geoboard activities in your classroom yet, your kids are going to  love  them!

5. Create a Shapes Photo Book (Pre-K Pages) – To introduce shape concepts to my son, I grabbed the book So Many Circles, So Many Squares from our library as the anchor to our learning ship. Using this book, we went on a shape scavenger hunt and made a fun shape keepsake.

6. Perfect Square Shapes Art (Pre-K Pages) – This Perfect Square art activity is so easy to set up and totally open-ended. It goes perfectly with the book and would be an excellent addition to a  shapes  unit.

7. Building Shapes with Craft Sticks (Pre-K Pages) – Pair this activity with the book  Shapes, Shapes, Shapes  by Tana Hoban and you’ve got the perfect low-prep shapes lesson!

8. Teaching 3D Shapes (Pre-K Pages) – Here are some of my favorite ideas for teaching 3D shapes to young children in pre-k or kindergarten. I also wrote some very simple 3D shape songs for you that incorporate hands on learning; keep reading to download the 3D shapes printable song charts.

9. Nature Shape Scavenger Hunt (Pre-K Pages) – A Star in My Orange  is a great way to reinforce shape recognition with your preschoolers. They will also immediately want to run outside for their very own shapes scavenger hunt in nature!

10. The Shape of Things Chalk Drawings (Pre-K Pages) – Shapes are found, identified, and drawn in all preschool classrooms! Discovering just how often circles, squares, and triangles occur in our everyday life make them relevant to children.

11. Shape Wands (Pre-K Pages) – Make some shape wands and turn your home or classroom into the perfect place for kids to learn and identify shapes.

12. Shape Exploration (Pre-K Pages) – After reading the wonderful book  Mouse Shapes  by Ellen Walsh , I thought it would be fun to make a mouse that could be used in a shape game. The mice in the book explore  shapes  so, why shouldn’t we?

13. Make a Tortilla Shape Snack (Pre-K Pages) – We have the perfect recipe for exploring a math concept. Read aloud one great children’s book. Make a healthy and yummy treat. Combine the two together and you have a lesson on  shapes .

14. Shapes Word Chart (Pre-K Pages) – This word chart focused on  shapes  but you can make a word chart for any topic.

15. “I Have Who Has” Shapes Game (Prekinders) – You may have seen the “I Have, Who Has” card games circulating the internet a lot lately, so this is a fun twist for Pre-K to teach shapes.

16. Games and Activities for Teaching Shapes (Prekinders) – Here are a fun few ways to teaching shapes, like shape bingo and a memory game.

17. Tracing Shapes on the Flannel Board (Teach Preschool) – A wonderful way to introduce letters and shapes while building pre-writing skills!

18. Hunting for Shapes (Teach Preschool) -Explore shapes with a fun and interactive game!

19. Exploring Shapes with Blocks on a Table Top (Teach Preschool) –  A simple and engaging exploration of shapes and colors!

20. Learning Shapes by Rolling a Ball (Hands On As We Grow) – Try a fun hands on activity for toddlers for a creative twist to learn shapes!

21. Finding Shapes at the Playground (Buggy and Buddy)- Just print out the free shape hunt printable and go searching for shapes at the playground with this fun geometry activity for children!

22. Geometric Shapes Math Activity (Little Bins for Little Hands) – This simple geometric shapes activity  for kids is easy to do at home or as a math center in school. It also makes a terrific STEAM project including a bit of art and design too.

23. Gruffalo Themed Shape Animals (Educators’ Spin on It) – Exploring shapes with young children can be such fun when you involve a few animal friends from The Gruffalo.

Shape Learning Activities for Preschoolers

24. Feed the Shape Monster Game (Imagination Tree) – Make a fun activity for preschoolers and school aged kids with this feed the hungry shape monsters sorting game!

25. Sticky Shape Bugs Activity (Mom Inspired Life) –   This was a great way to develop fine motor skills and critical thinking skills while learning about shapes.

26. Learning Shapes with Spaghetti Noodles (Teaching Mama) – Looking for a fun way to teach shapes? Well here’s a very fun way using spaghetti noodles! This hands-on activity also is a great sensory activity.

27. Matching Shapes to Outlines (Busy Toddler) – Create this fun easy DIY shape mat to practice shapes with your preschoolers.

28. Chalk Shapes Jumping Game (Craftulate) – All you need for this shape activity is some sidewalk chalk!

29. Open and Closed Polygons Game (JDaniel4’s Mom) Grab some LEGOs and have fun with these polygon games, like hockey!

30. Shape Sensory Squish Bag (Still Playing School) –  Create this sensory squish bag with triangles, circles, and squares. It’s irresistible to touch and talk about in the window or on the table. It’s super easy to make, too!

31. DIY Shape Puzzles (Munchkins and Moms) – If you have some Jenga blocks and markers, then this easy DIY shape puzzle will be a fun engaging activity for your preschoolers.

32. Stamping Shapes in Kinetic Sand (Still Playing School) – Stamping shapes into kinetic sand is a great opportunity to work on shape identification, count the sides and corners, and compare and contrast the shapes.

33. Making Trucks from Shapes (Powerful Mothering) – Using your wooden blocks to draw and create trucks!

34. DIY Waldorf Square (Rhythms of Play) – An easy DIY toy for kids made with wooden blocks and liquid watercolor paints.

35. Magazine Shape Hunt and Sort (Mom Inspired Life) – This magazine shape hunt is jam-packed with learning! Kids will learn shapes while they practice cutting, gluing and sorting. It’s also an awesome way to work on critical thinking and observation skills.

36. Building Rockets with Shapes (Stir the Wonder) – Building rockets with shapes is a fun way to review shapes and colors with toddlers and preschoolers!

37. Build on Shape Outlines (Brick by Brick) – Use wooden blocks in a new fun way and work on shapes at the same time!

38. Shapes in Our Neighborhood Book (Munchkins and Moms) – Go for a walk and look for shapes in the neighborhood and then create a photo book after!

39. Sorting Shapes in the Sensory Bin (Learning 4 Kids) – Your preschoolers will practice their shapes and fine motor skills while having fun with this shape sensory bin.

40. I Spy Shape Hunt (Munchkins and Moms) – Create these fun easy spy glasses and go on a shape hunt!

Shapes PInterest Board

2 thoughts on “Shapes Activities for Preschoolers”

' src=

I cant find these on your members site.

' src=

Hi Stacy, These are free so you can just download them right from Pre-K Pages. With more than 300 free printables at Pre-K Pages I didn’t want to clutter the membership vault with them.

Comments are closed.

shape based activity recognition

  • Skip to main content
  • Skip to primary sidebar

Fun with Mama

Toddler and preschool activities

Shapes Activities For Preschoolers

These hands-on 2D Shape activities for preschoolers will make learning shapes simple for teachers, and FUN for toddlers, preschoolers and kindergarteners. Did you know that learning shapes is a foundational skill? This can be done through having a preschool shapes theme week or by doing a shape of the week throughout the year.

In this post, I’m going to share with you my favorite shape resource that I’ve spent hours putting together so that you, my teacher friend, or fellow parent, don’t have to. I will also share my favorite free shape printables and simple activities you can do at home. You can check out my 2d shapes activity pack here.

shape activities for preschool

Please note that while these shape activities for kindergarteners and preschoolers can be used by various ages. Adult supervision is always required.

You can find our favorite kids activity pack supplies here.

colors and shapes activities for preschoolers and toddlers

This page will be updated constantly to add more shapes activities.

Did you know that preschool shapes recognition is an early math skill? You also may be wondering how to teach basic shapes to preschoolers or toddlers and what shapes preschoolers should learn.

The shapes preschoolers should know are:

Take it further by introducing more shapes:

With these activities, not only will they be learning basic preschool shapes they will also work on early math skills and even learn preschool colors too. We often sprinkle in some shape activities throughout our preschool themes and units because repetition is the key. Children are sponges and they learn through repetition. 

These activities will make teaching shapes more fun.

shape based activity recognition

The first few activities on this post are from our 2D Shapes Activity Pack

GET THE SHAPE ACTIVITY PACK HERE

shape based activity recognition

GET THE ABOVE SHAPE ACTIVITIES

2d shape posters.

shape based activity recognition

Yes, hang those posters up! Children are more observant than you realize, so while they are learning shapes it’s a great idea to hang up some shape posters in your home or classroom.

Shape 3 Part Cards

shape based activity recognition

3 part cards, also known as Nomenclature cards, refer to

  • The picture only card
  • The word only card
  • The whole card <— also the control card that includes the picture and word.

These Montessori inspired cards enrich vocabulary, and reading and writing skills. They help build connections. Students will match the correct words to the correct whole card.

Playdough Shapes Cards

shape based activity recognition

Children learn better when we can incorporate more of their senses. We prefer to use hands on activities that encourage children to move their hands and bodies. Using play dough to make shapes will not only help them learn how to form a shape, it will also strengthen their fine motor skills, strengthen their hands, and work on their visual perception.

2D Shape Sorting Mats

2d Shape Sorting Mats

These shape sorting mats include basic shapes and for a more advanced child, I’ve also included some real life shape objects. Children can sort the shapes to their correct mat. With a younger child you may opt to only give the student 2 mats at once, while with an older child you can have them sort shapes amongst all 12 mats.

2D Shape Roads

shape based activity recognition

Beep! Beep! Let’s have some fun with those toys at home. Learning can be fun, especially if you pair your favorite toy with a learning activity. Children will use a toy car to drive around the shape. Add more of a sensory aspect to it by adding shaving cream to the mat (make sure to laminate the mats first.) Have them drive through the messy shaving cream and then, when they are done, dump their cars in a water bin for a carwash. This adds a sensory aspect to it, and we know how much kids love Water Play

2D Shape Lacing Activity

shape based activity recognition

2D Craft Stick Shape Cards

shape based activity recognition

Children love to build and create. Not only is this great for working on problem solving skills and creativity, but they learn math skills that will help them in the future. Have students build shapes out of popsicle sticks. You can opt to add velcro dots to the edges of each stick to make the pieces more sticky. When done building shapes, see what else children opt to create.

Shape Count and Clip Cards Math Activity

shape based activity recognition

First, identify the shape and then students will count how many shapes are on the card. Then they will clip the correct number.

Match The Shape Buttons

shape based activity recognition

I have a 2.5 year old at home, and one of her favorite activities is anything that incorporates stickers or these shape buttons. Children will match the color and shape button to the corresponding spot on the mat. Do your little ones enjoy working with shape buttons too?

Sticker Shapes

shape based activity recognition

While we are working on my toddlers (or is she a preschooler)’s favorite activity type I incorporated the pom pom shape activity. I normally offer this activity two ways. The first is with pom poms. Children will place the correct pom pom color onto each shape’s white spot. Then, once we are done with pom poms I give her some circle stickers and she can place the stickers on the white circles.

2D Shape Fish Tank Hunt

shape based activity recognition

These fish come in different shapes and sizes. Students will identify the fish with the matching shape and then cover them up. You can use goldfish crackers to mark the shape.

2D Shape Puzzles

shape based activity recognition

For young children, cut the puzzle into two pieces. For older children, cut the shape into 4 pieces. Then have students match the shape to the shape word heading.

2D Shape Hunt Shape Mats

shape based activity recognition

The detective is on a mission. He is looking for specific shapes. Children will need to help the detective find the shape and then write the number they found. If you laminate the page children can use a dry erase marker to write the number. Or, if you don’t, you can also have them use a magnetic number instead.

Snail Shell 2D Shape Matching

Match the shape shell to the correct snail. You can make this a file folder game. Laminate the page and add velcro dots.

2D Shape Foldable Book

Fold the book then follow the prompts on each page to color, trace, and draw the shape.

 2D Shape Patterns

shape based activity recognition

Yep, there’s those buttons again. Like I mentioned before, we love them and if I’m going to purchase a material, I want to make sure we get to use them over and over again. These shape buttons pair perfectly with the shape pattern activity.

I am an Amazon affiliate, affiliate links are used in this post.

You can get the above fun shape activities here:

GET THE ABOVE SHAPE ACTIVITIES HERE

Shape bestsellers in the shop.

shape based activity recognition

More FUN Preschool shapes activities

Shapes themed toddler activities as well as preschool activities , constantly make a reappearance in our home. 

colors and shapes activities for preschoolers

More Of The Best Preschool Shapes Activities

Mystery shapes.

Check out two of my daughters favorite shapes activities in her new YouTube video… have your child learn along with her.

Preschool Shapes Matching Activity

shape based activity recognition

Make shapes collages using contact paper. This sticky shapes activity is a shapes sorting activity. The contact paper removes the need for glue and makes it easier for a toddler. This learning shapes and colors for toddler Printable activity looks as pretty as it is fun. My toddler had a blast placing the small stars in the big star.

Preschool Shapes Math Hunt Sensory Bin

shape based activity recognition

This puzzle shapes hunt activity works wonderfully with any puzzle you have at home. Kids will learn their shapes through a sensory experience. The Shapes sensory bin can be made harder by having your child close their eyes and feeling the shape. My toddler was impressed by my preschool shapes game.

Toddler Shapes Activity Fishing

shape based activity recognition

Preschool Shapes Paint Resist Activity

shape based activity recognition

Try out this free printable  shape scavenger hunt.

12 Preschool Shapes Sensory Bins from around the web

Shapes Themed sensory bins for toddlers and preschoolers

We love sensory bins in our home and these 12 shapes sensory bins from around the web give you many different options on teaching your child shapes through sensory exploration.

Jelly Beans Color and Preschool Shapes Matching

Who can resist colorful jelly beans play? This activity works on shape and color recognition.

Shapes Fine Motor Activity

shape based activity recognition

This activity is a little more advanced and works well for older preschoolers and elementary grade schoolers. I also enjoyed sewing and weaving with the kids. You can work on color recognition too.

Create a Preschool Shapes Stamp

Make shape stamps out of foam shapes and popsicle  sticks.

shape based activity recognition

Make a Shapes Canvas Painting

shape based activity recognition

Bingo Colors Printable Game

bingo colors printable game

  • Skip to primary navigation
  • Skip to main content
  • Skip to primary sidebar
  • MEMBER LOGIN

Hands On As We Grow®

Hands on kids activities for hands on moms. Focusing on kids activities perfect for toddlers and preschoolers.

12 Shape Activities for Toddlers! It’s Hip to be Square!

Math & 123s Toddlers Resources Shapes 24 Comments

These 12 hands on shape activities for toddlers (2 year olds) will make it hip to be square!

Toddlers will have fun learning shapes in a hands on way.

We’ve been having a little bit of fun lately doing some hands on shape activities!

George is on a shapes kick right now, anytime he asks for an activity, he’s been asking for ‘shapes’. I can’t say no to that!

I keep the shapes I work on with my toddlers in activities pretty basic.

Usually the square, triangle, circle.

And then I usually sneak in an odd one here and there just to introduce it, like the octagon (my kids seem to catch on to that one quickly because of stop signs), oval and rectangles.

These shape activities for toddlers are all very hands on and fun, the learning is just a plus!

How to Teach Shapes to Toddlers

First comes recognizing the shapes and then labeling them.

These shapes activities for toddlers are perfect for 2 year olds learning to recognize their shapes.

Building and making shapes is a plus for toddlers, as well as drawing and tracing them. This will come in time… preschoolers will be able to start doing this.

Find 50 activities that are perfect for your toddler  here.

Teaching shapes to toddlers can be a lot of fun with these simple hands on activities. Keep reinforcing what your teaching with a brand new activity!

12 shape activities for 2 year olds

These shape activities for toddlers are all very hands on and fun, the learning is just a plus!

I tend to stray away from printables and worksheets, so these activities pass my ‘hands-on’ learning test that all get toddlers very involved in learning shapes.

Shapes are fun to start spotting out in your every day lives too! Your two year old will start seeing wheels as circles, and houses as squares.

Make it a challenge when you’re out for a walk to see what shapes you can each spot!

Call them out and have a good time with it.

I bet you can even find an octagon (Stop signs!). That’s a great way to introduce a new shape like that.

For more learning activities for toddlers, check out these  14 activities that focus on shapes, letters, colors and numbers.

Is your toddler showing interest in shapes? Go ahead and jump on it with these super fun and simple activities that are play based learning.

There are also tons of super fun books about shapes for toddlers to reinforce what your learning with activities.

Shapes: My First Pop-up! (A Pop Magic Book) by Matthew Reinhart is a super fun one!

Download the FREE Week of Toddler Activities to do this week!

WANT TO SAVE THIS ACTIVITY?

Enter your email below & we'll send it straight to your inbox so you can access this activity later! Plus, you'll get simple activities from us every week!

  • Hidden Activity URL
  • Hidden Activity Title
  • Comments This field is for validation purposes and should be left unchanged.

About Jamie Reimer

Jamie learned to be a hands on mom by creating activities, crafts and art projects for her three boys to do. Jamie needed the creative outlet that activities provided to get through the early years of parenting with a smile! Follow Jamie on Pinterest and Instagram !

More Hands on Kids Activities to Try

shape based activity recognition

Reader Interactions

24 comments.

Remi Kate says

December 5, 2022 at 7:39 am

this article is very useful. l have a little brother and he is really likes such methods. now we are learning fruits and vegetables, this is not advertising, it’s just that this article may also be useful to someone https://wunderkiddy.com/category/vegetables-fruits-and-berries

carol klein says

April 30, 2020 at 10:04 am

Love the ideas. I have a story book of shapes That I read to my grand daughter. I write the word or drew the shape on a paper then I challenge her to find the name and shape in the book .. then she challenges me by drawing the shape of just writing a letter she knows and I have to find them..

swetha says

April 24, 2016 at 1:36 pm

this will be very useful to me as a teacher.

sutton says

November 16, 2015 at 8:17 am

I am a home base teacher in Cleveland, ohio and the activities that we were using was boring for the kids. so I found your website which gave me and the children more learning time with the same activity and I have more questions to ask the children during work time. I look on your website everyday for some thing new and different.

Bonnie says

October 22, 2015 at 12:16 pm

I love your ideas with shapes. I decided to trace my small colored blocks onto a poster board then I had my toddlers match the shape on the board to the blocks in the basket. They loved it. Thanks for sharing your great ideas. Have a blessed day.

August 29, 2013 at 6:54 am

Such great ideas! I love the driveway shape maze! In kindergarten (so a little bit older wee ones) we would play ‘guess my shape’. I would start drawing a shape (one angled line for example) and the kids would guess the shape just from that first clue, then I’d add a second line, etc. until they all got the shape. It was great in helping them look at shapes more closely – what makes a triangle different from other shapes? Later in the year the kids would play this game on their own with friends. Love the hands on learning!! Excited to play chalk maze !

Leave a Comment Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed .

shape based activity recognition

What Parents Have to Say…

Shop ebooks of activities.

Activity Capsule Grab & Go Bags

Get activity plans delivered to your inbox, every week!

Activities that hands-on parents absolutely love.

shape based activity recognition

Why Routines for Kids are Important

shape based activity recognition

Improve Kids Fine Motor Skills with 30 Materials & Activities

Dive into early learning with 40+ number activities for preschoolers!

40+ Awesome Number Activities for Preschoolers

shape based activity recognition

50+ Simple Activities for Toddlers

30 gross motor activities for preschooler that are on the go all the time.

Gross Motor Activities for Preschoolers: The Top 35!

shape based activity recognition

How to Make a Lava Lamp Experiment Without Alka Seltzer

Get started having fun with your kids.

PLAN THE FUN WITH THE FREE KIDS ACTIVITIES PLANNER! AND RECEIVE ACTIVITIES EVERY WEEK!

Hands On As We Grow®

  • Preschoolers
  • Kindergartners
  • Grade School
  • Literacy & ABCs
  • Math & 123s
  • Art Projects
  • Gross Motor
  • Shop Activity Plans
  • Member Login

Book cover

Contactless Human Activity Analysis pp 43–81 Cite as

Skeleton-Based Activity Recognition: Preprocessing and Approaches

  • Sujan Sarker 6 ,
  • Sejuti Rahman 6 ,
  • Tonmoy Hossain 7 ,
  • Syeda Faiza Ahmed 6 ,
  • Lafifa Jamal 6 &
  • Md Atiqur Rahman Ahad 8 , 9  
  • First Online: 24 March 2021

870 Accesses

7 Citations

Part of the Intelligent Systems Reference Library book series (ISRL,volume 200)

Research in Activity Recognition is one of the thriving areas in the field of computer vision. This development comes into existence by introducing the skeleton-based architectures for action recognition and related research areas. By advancing the research into real-time scenarios, practitioners find it fascinating and challenging to work on human action recognition because of the following core aspects—numerous types of distinct actions, variations in the multimodal datasets, feature extraction, and view adaptiveness. Moreover, hand-crafted features and depth sequence models cannot perform efficiently on the multimodal representations. Consequently, recognizing many action classes by extracting some smart and discriminative features is a daunting task. As a result, deep learning models are adapted to work in the field of skeleton-based action recognition. This chapter entails all the fundamental aspects of skeleton-based action recognition, such as—skeleton tracking, representation, preprocessing techniques, feature extraction, and recognition methods. This chapter can be a beginning point for a researcher who wishes to work in action analysis or recognition based on skeleton joint-points.

This is a preview of subscription content, log in via an institution .

Buying options

  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
  • Available as EPUB and PDF
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info
  • Durable hardcover edition

Tax calculation will be finalised at checkout

Purchases are for personal use only

Baek, S., Kwang, I.K., Kim, T.-K.: Augmented skeleton space transfer for depth-based hand pose estimation. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 8330–8339 (2018)

Google Scholar  

Mehta, D., Sridhar, S., Sotnychenko, O., Rhodin, H., Shafiei, M., Seidel, H.-P., Weipeng, X., Casas, D., Theobalt, C.: Vnect: Real-time 3d human pose estimation with a single rgb camera. ACM Trans. Graph. (TOG) 36 (4), 1–14 (2017)

Article   Google Scholar  

Ling, J., Tian, L., Li, C.: 3d human activity recognition using skeletal data from rgbd sensors. In: International Symposium on Visual Computing, pp. 133–142. Springer (2016)

Balakrishnan, S., Rice, J.M., Walker, S.H., Carroll, A.S., Dow-Hygelund, C.C., Goodwin, A.K., Mullin, J.M., Rattenbury, T.L., Rooke-Ley, J.M., Schmitt, J.M., et al.: Action detection and activity classification, May 31 2016. US Patent 9,352,207

Wang, J., Liu, Z., Ying, W., Yuan, J.: Learning actionlet ensemble for 3d human action recognition. IEEE Trans. Pattern Analy. Mach. Intell. 36 (5), 914–927 (2013)

Wang, L., Gu, T., Tao, X., Lu, J.: Sensor-based human activity recognition in a multi-user scenario. In: European Conference on Ambient Intelligence, pp. 78–87. Springer (2009)

Batabyal, T., Chattopadhyay, T., Mukherjee, D.P.: Action recognition using joint coordinates of 3d skeleton data. In: 2015 IEEE International Conference on Image Processing (ICIP), pp. 4107–4111. IEEE (2015)

Kong, Y., Fu, Y.: Bilinear heterogeneous information machine for rgb-d action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1054–1062 (2015)

Seidenari, L., Varano, C., Berretti, S., Bimbo, A., Pala, P.: Recognizing actions from depth cameras as weakly aligned multi-part bag-of-poses. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp. 479–485 (2013)

Pham, H.-H., Khoudour, L., Crouzil, A., Zegers, P., Velastin, S.A.: Exploiting deep residual networks for human action recognition from skeletal data. Comput. Vis. Image Underst. 170 , 51–66 (2018)

Presti, L.L., Cascia, M.L.: 3d skeleton-based human action classification: a survey. Pattern Recogn. 53 , 130–147 (2016)

Chen, Y., Tian, Y., He, M.: Monocular human pose estimation: a survey of deep learning-based methods. Comput. Vis. Image Underst. 192 , 102897, 03 (2020)

Zhang, A., Ma, X., Song, R., Rong, X., Tian, X., Tian, G., Li, Y.: Deep learning based human action recognition: a survey. In: 2017 Chinese Automation Congress (CAC), pp. 3780–3785. IEEE (2017)

Asadi-Aghbolaghi, M., Clapés, A., Bellantonio, M., Escalante, H.J., Ponce-López, V., Baró, X., Guyon, I., Kasaei, S., Escalera, S.: Deep learning for action and gesture recognition in image sequences: a survey. In: Gesture Recognition, pp. 539–578. Springer (2017)

Wang, L., Huynh, D.Q., Koniusz, D.Q.: A comparative review of recent kinect-based action recognition algorithms. IEEE Trans. Image Process. 29 , 15–28 (2019)

Article   MathSciNet   Google Scholar  

Jegham, I., Khalifa, A.B., Alouani, I., Mahjoub, M.A.: Vision-based human action recognition: an overview and real world challenges. Forensic Sci. Int.: Digital Investig. 32 , 200901 (2020)

Cao, Z., Simon, T., Wei, S.-E., Sheikh, S.-E.: Realtime multi-person 2d pose estimation using part affinity fields. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7291–7299 (2017)

Simon, T., Joo, H., Matthews, I., Sheikh, Y.: Hand keypoint detection in single images using multiview bootstrapping. In: CVPR (2017)

Wei, S.-E., Ramakrishna, S.-E., Kanade, T., Sheikh. Y.: Convolutional pose machines. In: CVPR (2016)

Rahmani, H., Mian, A.: 3d action recognition from novel viewpoints. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1506–1515 (2016)

Vieira, A.W., Nascimento, E.R., Oliveira, G.L., Liu, Z., Campos, M.F.M.: Stop: space-time occupancy patterns for 3d action recognition from depth map sequences. In: Iberoamerican Congress on Pattern Recognition, pp. 252–259. Springer (2012)

Cavazza, J., Zunino, A., San Biagio, M., Murino, V.: Kernelized covariance for action recognition. In: 2016 23rd International Conference on Pattern Recognition (ICPR), pp. 408–413. IEEE (2016)

Materzynska, J., Xiao, J., Herzig, R., Xu, H., Wang, X., Darrell, T.: Something-else: compositional action recognition with spatial-temporal interaction networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1049–1059 (2020)

Yang, J., Liu, Wu, Yuan, J.: Mei, T: Hierarchical soft quantization for skeleton-based human action recognition. IEEE Trans, Multimedia (2020)

Huang, J., Xiang, X., Gong, X., Zhang, B., et al.: Long-short graph memory network for skeleton-based action recognition. In: The IEEE Winter Conference on Applications of Computer Vision, pp. 645–652 (2020)

Si, C., Chen, W., Wang, W., Wang, L., Tan, T.: An attention enhanced graph convolutional lstm network for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1227–1236 (2019)

Yan, S., Li, Z., Xiong, Y., Yan, H., Lin, D.: Convolutional sequence generation for skeleton-based action synthesis. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 4394–4402 (2019)

Zhao, R., Wang, K., Su, K., Ji, Q.: Bayesian graph convolution lstm for skeleton based action recognition. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 6882–6892 (2019)

Shi, L., Zhang, Y., Cheng, J., Lu, H.: Two-stream adaptive graph convolutional networks for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 12026–12035 (2019)

Yang, H., Yan, D., Zhang, L., Li, D., Sun, Y.D., You, S.D., Maybank, S.J.: Feedback graph convolutional network for skeleton-based action recognition. arXiv preprint arXiv:2003.07564 (2020)

Zhu, G., Zhang, L., Li, H., Shen, P., Afaq Ali Shah, S., Bennamoun, M.: Topology-learnable graph convolution for skeleton-based action recognition. Pattern Recogn. Lett. (2020)

Chen, Y., Ma, G., Yuan, C., Li, B., Zhang, H., Wang, F., Hu, W.: Graph convolutional network with structure pooling and joint-wise channel attention for action recognition. Pattern Recogn., p. 107321 (2020)

Huang, L., Huang, Y., Ouyang, W., Wang, L. et al.: Part-level graph convolutional network for skeleton-based action recognition (2020)

Tang, Y., Tian, Y., Lu, J., Li, P., Zhou, J.: Deep progressive reinforcement learning for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5323–5332 (2018)

Caetano, C., Brémond, F., Schwartz, W.R.: Skeleton image representation for 3d action recognition based on tree structure and reference joints. In: 2019 32nd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI), pp. 16–23. IEEE (2019)

Ke, Q., Bennamoun, M., An, A., Sohel, F., Boussaid, F.: A new representation of skeleton sequences for 3d action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3288–3297 (2017)

Liliana [Lo Presti], Marco [La Cascia]: 3d skeleton-based human action classification: a survey. Pattern Recogn. 53 , 130–147 (2016)

Chaudhry, R., Ofli, F., Kurillo, G., Bajcsy, R., Vidal, R.: Bio-inspired dynamic 3d discriminative skeletal features for human action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, pp. 471–478 (2013)

Slama, R., Wannous, H., Daoudi, M., Srivastava, A.: Accurate 3d action recognition using learning on the grassmann manifold. Pattern Recogn. 48 (2), 556–567 (2015)

Li, X., Zhang, Y., Zhang, J.: Improved key poses model for skeleton-based action recognition. In: Pacific Rim Conference on Multimedia, pp. 358–367. Springer (2017)

Cai, L., Liu, C., Yuan, R., Ding, H.: Human action recognition using lie group features and convolutional neural networks. Nonlinear Dyn., pp. 1–11 (2020)

Ghorbel, E., Demisse, G., Aouada, D., Ottersten, B.: Fast adaptive reparametrization (far) with application to human action recognition. IEEE Signal Process. Lett. 27 , 580–584 (2020)

Huang, Z., Wan, C., Probst, T., Van Gool, L.: Deep learning on lie groups for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 6099–6108 (2017)

de Boissiere, A.M., Noumeir, R.: Infrared and 3d skeleton feature fusion for rgb-d action recognition. arXiv preprint arXiv:2002.12886 (2020)

Lee, I., Kim, D., Kang, S., Lee, S.: Ensemble deep learning for skeleton-based action recognition using temporal sliding lstm networks. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1012–1020 (2017)

Rahmani, H., Bennamoun, M.: Learning action recognition model from depth and skeleton videos. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 5832–5841 (2017)

Zhang, P., Lan, C., Xing, J., Zeng, W., Xue, J., Zheng, N.: View adaptive recurrent neural networks for high performance human action recognition from skeleton data. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 2117–2126 (2017)

Li, R., Fu, H., Lo, W., Chi, Z., Song, Z., Wen, D.: Skeleton-based action recognition with key-segment descriptor and temporal step matrix model. IEEE Access 7 , 169782–169795 (2019)

Rahmani, H., Bennamoun, M.: Learning action recognition model from depth and skeleton videos. In: 2017 IEEE International Conference on Computer Vision (ICCV), pp. 5833–5842 (2017)

Nie, Q., Wang, J., Wang, X., Liu, Y.: View-invariant human action recognition based on a 3d bio-constrained skeleton model. IEEE Trans. Image Process. 28 (8), 3959–3972 (2019)

Li, S., Jiang, T., Tian, Y., Huang, T.: 3d human skeleton data compression for action recognition. In: 2019 IEEE Visual Communications and Image Processing (VCIP), pp. 1–4 (2019)

Nie, W., Wang, W., Huang, X.: Srnet: Structured relevance feature learning network from skeleton data for human action recognition. IEEE Access 7 , 132161–132172 (2019)

Shi, L., Zhang, Y., Cheng, J., Lu, H.: Skeleton-based action recognition with directed graph neural networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 7912–7921 (2019)

Li, S., Jiang, T., Huang, T., Tian, Y.: Global co-occurrence feature learning and active coordinate system conversion for skeleton-based action recognition. In: The IEEE Winter Conference on Applications of Computer Vision, pp. 586–594 (2020)

Du, Y., Wang, W., Wang, L.: Hierarchical recurrent neural network for skeleton based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1110–1118 (2015)

Su, K., Liu, X., Shlizerman, E., Predict & cluster: Unsupervised skeleton based action recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9631–9640 (2020)

Zhang, P., Lan, C., Zeng, W., Xing, J., Xue, J., Zheng, N.: Semantics-guided neural networks for efficient skeleton-based human action recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1112–1121 (2020)

Raj, B.N., Subramanian, A., Ravichandran, K., Venkateswaran, N.: Exploring techniques to improve activity recognition using human pose skeletons. In: Proceedings of the IEEE Winter Conference on Applications of Computer Vision Workshops, pp. 165–172 (2020)

Huang, J., Huang, Z., Xiang, X., Gong, X., Zhang, B.: Long-short graph memory network for skeleton-based action recognition. In: The IEEE Winter Conference on Applications of Computer Vision (WACV), March 2020

Huynh, D.Q.: Metrics for 3d rotations: Comparison and analysis. J. Math. Imaging Vis. 35 (2), 155–164 (2009)

Zhu, W., Lan, C., Xing, J., Zeng, W., Li, Y., Shen, L., Xie, X.: Co-occurrence feature learning for skeleton based action recognition using regularized deep lstm networks. In: Thirtieth AAAI Conference on Artificial Intelligence (2016)

Morais, R., Le, V., Tran, T., Saha, B., Mansour, M., Venkatesh, S.: Learning regularity in skeleton trajectories for anomaly detection in videos. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 11996–12004 (2019)

Gaglio, S., Re, G.L., Morana, M.: Human activity recognition process using 3-d posture data. IEEE Trans. Human-Mach. Syst. 45 (5), 586–597 (2014)

Naveenkumar, M., Domnic, S.: Skeleton joint difference maps for 3d action recognition with convolutional neural networks. In: International Conference on Recent Trends in Image Processing and Pattern Recognition, pp. 144–150. Springer (2018)

Wang, P., Li, W., Gao, Z., Zhang, J., Tang, C., Ogunbona, P.O.: Action recognition from depth maps using deep convolutional neural networks. IEEE Trans. Human-Mach. Syst. 46 (4), 498–509 (2015)

Yang, X., Zhang, C., Tian, Y.L.: Recognizing actions using depth motion maps-based histograms of oriented gradients. In: Proceedings of the 20th ACM International Conference on Multimedia, pp. 1057–1060 (2012)

Li, B., Dai, Y., Cheng, X., Chen, H., Lin, Y., He, M.: Skeleton based action recognition using translation-scale invariant image mapping and multi-scale deep cnn. In: 2017 IEEE International Conference on Multimedia & Expo Workshops (ICMEW), pp. 601–604. IEEE (2017)

Huynh-The, T., Hua, C.-H., Tu, N.A., Kim, J.-W., Kim, S.-H., Kim, D.-S.: 3d action recognition exploiting hierarchical deep feature fusion model. In: 2020 14th International Conference on Ubiquitous Information Management and Communication (IMCOM), pp. 1–3. IEEE (2020)

Li, M., Chen, S., Chen, X., Zhang, Y., Wang, Y., Tian, Q.: Actional-structural graph convolutional networks for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3595–3603 (2019)

Liu, J., Liu, Y., Wang, Y., Prinet, V., Xiang, S., Pan, C.: Decoupled representation learning for skeleton-based gesture recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 5751–5760 (2020)

Boser, B.E., Guyon, I.M., Vapnik, V.N.: A training algorithm for optimal margin classifiers. In: Proceedings of the Fifth annual Workshop on Computational Learning Theory, pp. 144–152 (1992)

Cortes, Corinna, Vapnik, Vladimir: Support-vector networks. Mach. Learn. 20 (3), 273–297 (1995)

MATH   Google Scholar  

Vemulapalli, R., Arrate, F., Chellappa, R.: Human action recognition by representing 3d skeletons as points in a lie group. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 588–595 (2014)

Boulahia, S.Y., Anquetil, E., Kulpa, R., Multon, F.: Hif3d: Handwriting-inspired features for 3d skeleton-based action recognition. In: 2016 23rd International Conference on Pattern Recognition (ICPR), pp. 985–990. IEEE (2016)

Li, X., Zhang, Y., Liao, D.: Mining key skeleton poses with latent svm for action recognition. Appl. Comput. Intell. Soft Comput. (2017)

Xu, D., Xiao, X., Wang, X., Wang, J.: Human action recognition based on kinect and pso-svm by representing 3d skeletons as points in lie group. In: 2016 International Conference on Audio, Language and Image Processing (ICALIP), pp. 568–573. IEEE (2016)

Liu, M., He, Q., Liu, H.: Fusing shape and motion matrices for view invariant action recognition using 3d skeletons. In: 2017 IEEE International Conference on Image Processing (ICIP), pp. 3670–3674. IEEE (2017)

Weng, J., Weng, C., Yuan, J.: Spatio-temporal naive-bayes nearest-neighbor (st-nbnn) for skeleton-based action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4171–4180 (2017)

Tang, N.C., Lin, Y.-Y., Hua, J.-H., Weng, M.-F., Mark Liao, H.-Y.: Human action recognition using associated depth and skeleton information. In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 4608–4612. IEEE (2014)

Ubalde, S., Gómez-Fernández, F., Goussies, N.A., Mejail, M.: Skeleton-based action recognition using citation-knn on bags of time-stamped pose descriptors. In: 2016 IEEE International Conference on Image Processing (ICIP), pp. 3051–3055. IEEE (2016)

Li, Y., Guo, T., Xia, R., Liu, X.: A novel skeleton spatial pyramid model for skeleton-based action recognition. In: 2019 IEEE 4th International Conference on Signal and Image Processing (ICSIP), pp. 16–20. IEEE (2019)

Liu, Z., Zhang, C., Tian, Y.: 3d-based deep convolutional neural network for action recognition with depth sequences. Image Vis. Comput. 55 , 93–100 (2016)

Wang, H., Wang, L.: Beyond joints: Learning representations from primitive geometries for skeleton-based action recognition and detection. IEEE Trans. Image Process. 27 (9), 4382–4394 (2018)

Zhang, P., Lan, C., Xing, J., Zeng, W., Xue, J., Zheng, N.: View adaptive neural networks for high performance skeleton-based human action recognition. IEEE Trans. Pattern Anal. Mach. Intell. 41 (8), 1963–1978 (2019)

Si, C., Jing, Y., Wang, W., Wang, L., Tan, T.: Skeleton-based action recognition with hierarchical spatial reasoning and temporal stack learning network. Pattern Recogn., p. 107511 (2020)

Yang, D., Li, M.M., Fu, H., Fan, J., Leung, H.: Centrality graph convolutional networks for skeleton-based action recognition. arXiv preprint arXiv:2003.03007 (2020)

Ke, Q., Bennamoun, M., An, S., Sohel, F., Boussaid, F.: Learning clip representations for skeleton-based 3d action recognition. IEEE Trans. Image Process. 27 (6), 2842–2855 (2018)

Tian, D., Lu, Z.-M., Chen, X., Ma, L.-H.: An attentional spatial temporal graph convolutional network with co-occurrence feature learning for action recognition. Multimedia Tools Appl., 1–19 (2020)

Liu, A.-A., Yu-Ting, S., Jia, P.-P., Gao, Z., Hao, T., Yang, Z.-X.: Multiple/single-view human action recognition via part-induced multitask structural learning. IEEE Trans. Cybern. 45 (6), 1194–1208 (2014)

Yang, Y., Deng, C., Tao, D., Zhang, S., Liu, W., Gao, X.: Latent max-margin multitask learning with skelets for 3-d action recognition. IEEE Trans. Cybern. 47 (2), 439–448 (2016)

Nguyen, X.S., Brun, L., Lézoray, O., Bougleux, S.: A neural network based on spd manifold learning for skeleton-based hand gesture recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 12036–12045 (2019)

Zhang, T., Zheng, W., Cui, Z., Zong, Y., Li, C., Zhou, X., Yang, J.: Deep manifold-to-manifold transforming network for skeleton-based action recognition. IEEE Trans, Multimedia (2020)

Book   Google Scholar  

Devanne, M., Wannous, H., Berretti, S., Pala, P., Daoudi, M., Bimbo, D.: Alberto: 3-d human action recognition by shape analysis of motion trajectories on riemannian manifold. IEEE Trans. Cybern. 45 (7), 1340–1352 (2014)

Li, J., Xie, X., Pan, Q., Cao, Y., Zhao, Z., Shi, G.: Sgm-net: Skeleton-guided multimodal network for action recognition. Pattern Recogn., 107356 (2020)

Liu, J., Wang, G., Duan, L.-Y., Abdiyeva, K., Kot, A.C.: Skeleton-based human action recognition with global context-aware attention lstm networks. IEEE Trans. Image Process. 27 (4), 1586–1599 (2017)

Zheng, W., Li, L., Zhang, Z., Huang, Y., Wang, L.: Relational network for skeleton-based action recognition. In: 2019 IEEE International Conference on Multimedia and Expo (ICME), pp. 826–831. IEEE (2019)

Mahasseni, B., Todorovic, S.: Regularizing long short term memory with 3d human-skeleton sequences for action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 3054–3062 (2016)

Han, Y., Chung, S.-L., Ambikapathi, A., Chan, J.-S., Lin, W.-Y., Su, S.-F.: Robust human action recognition using global spatial-temporal attention for human skeleton data. In: 2018 International Joint Conference on Neural Networks (IJCNN), pp. 1–8. IEEE (2018)

Song, S., Lan, C., Xing, J., Zeng, W., Jiaying, L.: Spatio-temporal attention-based lstm networks for 3d action recognition and detection. IEEE Trans. Image Process. 27 (7), 3459–3471 (2018)

Cheng, K., Zhang, Y., He, X., Chen, W., Cheng, J., Lu, H.: Skeleton-based action recognition with shift graph convolutional network. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 183–192 (2020)

Li, M., Chen, S., Zhao, Y., Zhang, Y., Wang, Y., Tian, Q.: Dynamic multiscale graph neural networks for 3d skeleton based human motion prediction. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 214–223 (2020)

Liu, Z., Zhang, H., Chen, Z., Wang, Z., Ouyang, W.: Disentangling and unifying graph convolutions for skeleton-based action recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 143–152 (2020)

Gao, X., Li, K., Zhang, Y., Miao, Q., Sheng, L., Xie, J., Xu, J.: 3d skeleton-based video action recognition by graph convolution network. In: 2019 IEEE International Conference on Smart Internet of Things (SmartIoT), pp. 500–501. IEEE (2019)

Li, C., Cui, Z., Zheng, W., Chunyan, X., Ji, R., Yang, J.: Action-attending graphic neural network. IEEE Trans. Image Process. 27 (7), 3657–3670 (2018)

Song, Y.-F., Zhang, Z., Wang, L.: Richly activated graph convolutional network for action recognition with incomplete skeletons. In: 2019 IEEE International Conference on Image Processing (ICIP), pp. 1–5. IEEE (2019)

Ye, F., Tang, H., Wang, X., Liang, X.: Joints relation inference network for skeleton-based action recognition. In: 2019 IEEE International Conference on Image Processing (ICIP), pp. 16–20. IEEE (2019)

Zhang, X., Xu, C., Tao, D.: Context aware graph convolution for skeleton-based action recognition. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 14333–14342 (2020)

Zhang, G., Zhang, X.: Multi-heads attention graph convolutional networks for skeleton-based action recognition. In: 2019 IEEE Visual Communications and Image Processing (VCIP), pp. 1–4. IEEE (2019)

Si, C., Jing, Y., Wang, W., Wang, L., Tan, T.: Skeleton-based action recognition with spatial reasoning and temporal stack learning. In: Proceedings of the European Conference on Computer Vision (ECCV), pp. 103–118 (2018)

Zare, A., Moghaddam, H.A., Sharifi, A.: Video spatiotemporal mapping for human action recognition by convolutional neural network. Pattern Anal. Appl. 23 (1), 265–279 (2020)

Cho, S., Maqbool, M., Liu, F., Foroosh, H.: Self-attention network for skeleton-based human action recognition. In: The IEEE Winter Conference on Applications of Computer Vision, pp. 635–644 (2020)

Jiang, M., Pan, N., Kong, J.: Spatial-temporal saliency action mask attention network for action recognition. J. Vis. Commun. Image Represent., p. 102846 (2020)

Yang, Z., Li, Y., Yang, J., Luo, J.: Action recognition with spatio-temporal visual attention on skeleton image sequences. IEEE Trans. Circuits Syst. Video Technol. 29 (8), 2405–2415 (2019)

Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., Natsev, P. et al.: The kinetics human action video dataset. arXiv preprint arXiv:1705.06950 (2017)

Shahroudy, A., Liu, J., Ng, T.-T., Wang, G.: Ntu rgb+d: a large scale dataset for 3d human activity analysis. In: IEEE Conference on Computer Vision and Pattern Recognition, June 2016

Liu, J., Shahroudy, A., Perez, M., Wang, G., Duan, L.-Y., Kot, A.C.: Ntu rgb+d 120: a large-scale benchmark for 3d human activity understanding. IEEE Trans. Pattern Anal. Mach, Intell (2019)

Yan, S., Xiong, Y., Lin, D.: Spatial temporal graph convolutional networks for skeleton-based action recognition. In: Thirty-Second AAAI Conference on Artificial Intelligence (2018)

Carreira, J., Zisserman, A.: Quo vadis, action recognition? a new model and the kinetics dataset. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 6299–6308 (2017)

Dong, J., Gao, Y., Lee, H.J., Zhou, H., Yao, Y., Fang, Z., Huang, B.: Action recognition based on the fusion of graph convolutional networks with high order features. Appl. Sci. 10 (4), 1482 (2020)

Liu, J., Wang, G., Hu, P., Duan, L.-Y., Kot, A.C.: Global context-aware attention lstm networks for 3d action recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1647–1656 (2017)

Liu, M., Liu, H., Chen, C.: Enhanced skeleton visualization for view invariant human action recognition. Pattern Recogn. 68 , 346–362 (2017)

Jian-Fang, H., Zheng, W.-S., Ma, L., Wang, G., Lai, J., Zhang, J.: Early action prediction by soft regression. IEEE Trans. Pattern Ana. Mach. Intell. 41 (11), 2568–2583 (2018)

Liu, J., Shahroudy, A., Wang, G., Duan, L.-Y., Kot, A.C.: Skeleton-based online action prediction using scale selection network. IEEE Trans. Pattern Anal. Mach. Intell. 42 (6), 1453–1467 (2019)

Papadopoulos, K., Ghorbel, E., Aouada, D., Ottersten, B.: Vertex feature encoding and hierarchical temporal modeling in a spatial-temporal graph convolutional network for action recognition. arXiv preprint arXiv:1912.09745 , 2019

Huynh-The, T., Hua, C.-H., Tu, N.A., Kim, D.-S.: Learning geometric features with dual–stream cnn for 3d action recognition. In: ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2353–2357. IEEE (2020)

Zhang, X., Xu, C., Tian, X., Tao, D.: Graph edge convolutional neural networks for skeleton-based action recognition. IEEE Trans. Neural Networks Learn, Syst (2019)

Li, B., Li, X., Zhang, Z., Fei, W.: Spatio-temporal graph routing for skeleton-based action recognition. Proceedings of the AAAI Conference on Artificial Intelligence 33 , 8561–8568 (2019)

Download references

Author information

Authors and affiliations.

Department of Robotics and Mechatronics Engineering, University of Dhaka, Dhaka, Bangladesh

Sujan Sarker, Sejuti Rahman, Syeda Faiza Ahmed & Lafifa Jamal

Department of Computer Science and Engineering, Ahsanullah University of Science and Technology, Dhaka, Bangladesh

Tonmoy Hossain

Department of Electrical and Electronic Engineering, University of Dhaka, Dhaka, Bangladesh

Md Atiqur Rahman Ahad

Department of Media Intelligent, Osaka University, Osaka, Japan

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Sujan Sarker .

Editor information

Editors and affiliations.

Multimedia R&D and Standards, Qualcomm Technologies Inc., San Diego, CA, USA

Upal Mahbub

College of Information and Computer Sciences, University of Massachusetts Amherst, Amherst, MA, USA

Tauhidur Rahman

Rights and permissions

Reprints and permissions

Copyright information

© 2021 Springer Nature Switzerland AG

About this chapter

Cite this chapter.

Sarker, S., Rahman, S., Hossain, T., Faiza Ahmed, S., Jamal, L., Ahad, M.A.R. (2021). Skeleton-Based Activity Recognition: Preprocessing and Approaches. In: Ahad, M.A.R., Mahbub, U., Rahman, T. (eds) Contactless Human Activity Analysis. Intelligent Systems Reference Library, vol 200. Springer, Cham. https://doi.org/10.1007/978-3-030-68590-4_2

Download citation

DOI : https://doi.org/10.1007/978-3-030-68590-4_2

Published : 24 March 2021

Publisher Name : Springer, Cham

Print ISBN : 978-3-030-68589-8

Online ISBN : 978-3-030-68590-4

eBook Packages : Intelligent Technologies and Robotics Intelligent Technologies and Robotics (R0)

Share this chapter

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Publish with us

Policies and ethics

  • Find a journal
  • Track your research

U.S. flag

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

  • Publications
  • Account settings
  • Advanced Search
  • Journal List
  • J Healthc Eng
  • v.2017; 2017

Logo of jhe

A Review on Human Activity Recognition Using Vision-Based Method

Shugang zhang.

1 College of Information Science and Engineering, Ocean University of China, Qingdao, China

Zhiqiang Wei

2 Department of Computer Science and Technology, Tsinghua University, Beijing, China

Shuang Wang

Human activity recognition (HAR) aims to recognize activities from a series of observations on the actions of subjects and the environmental conditions. The vision-based HAR research is the basis of many applications including video surveillance, health care, and human-computer interaction (HCI). This review highlights the advances of state-of-the-art activity recognition approaches, especially for the activity representation and classification methods. For the representation methods, we sort out a chronological research trajectory from global representations to local representations, and recent depth-based representations. For the classification methods, we conform to the categorization of template-based methods, discriminative models, and generative models and review several prevalent methods. Next, representative and available datasets are introduced. Aiming to provide an overview of those methods and a convenient way of comparing them, we classify existing literatures with a detailed taxonomy including representation and classification methods, as well as the datasets they used. Finally, we investigate the directions for future research.

1. Introduction

Human activity recognition (HAR) is a widely studied computer vision problem. Applications of HAR include video surveillance, health care, and human-computer interaction. As the imaging technique advances and the camera device upgrades, novel approaches for HAR constantly emerge. This review aims to provide a comprehensive introduction to the video-based human activity recognition, giving an overview of various approaches as well as their evolutions by covering both the representative classical literatures and the state-of-the-art approaches.

Human activities have an inherent hierarchical structure that indicates the different levels of it, which can be considered as a three-level categorization. First, for the bottom level, there is an atomic element and these action primitives constitute more complex human activities. After the action primitive level, the action/activity comes as the second level. Finally, the complex interactions form the top level, which refers to the human activities that involve more than two persons and objects. In this paper, we follow this three-level categorization namely action primitives, actions/activities, and interactions. This three-level categorization varies a little from previous surveys [ 1 – 4 ] and maintains a consistent theme. Action primitives are those atomic actions at the limb level, such as “stretching the left arm,” and “raising the right leg.” Atomic actions are performed by a specific part of the human body, such as the hands, arms, or upper body part [ 4 ]. Actions and activities are used interchangeably in this review, referring to the whole-body movements composed of several action primitives in temporal sequential order and performed by a single person with no more person or additional objects. Specifically, we refer the terminology human activities as all movements of the three layers and the activities/actions as the middle level of human activities. Human activities like walking, running, and waving hands are categorized in the actions/activities level. Finally, similar to Aggarwal et al.'s review [ 2 ], interactions are human activities that involve two or more persons and objects. The additional person or object is an important characteristic of interaction. Typical examples of interactions are cooking which involves one person and various pots and pans and kissing that is performed by two persons.

This review highlights the advances of image representation approaches and classification methods in vision-based activity recognition. Generally, for representation approaches, related literatures follow a research trajectory of global representations, local representations, and recent depth-based representations ( Figure 1 ). Earlier studies attempted to model the whole images or silhouettes and represent human activities in a global manner. The approach in [ 5 ] is an example of global representation in which space-time shapes are generated as the image descriptors. Then, the emergency of space-time interest points (STIPs) proposed in [ 6 ] triggered significant attention to a new local representation view that focuses on the informative interest points. Meanwhile, local descriptors such as histogram of oriented gradients (HOG) and histogram of optical flow (HOF) oriented from object recognition are widely used or extended to 3D in HAR area. With the upgrades of camera devices, especially the launch of RGBD cameras in the year 2010, depth image-based representations have been a new research topic and have drawn growing concern in recent years.

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.001.jpg

Research trajectory of activity representation approaches.

On the other hand, classification techniques keep developing in step with machine learning methods. In fact, lots of classification methods were not originally designed for HAR. For instance, dynamic time warping (DTW) and hidden Markov model (HMM) were first used in speech recognition [ 7 , 8 ], while the recent deep learning method is first developed for large amount image classification [ 9 ]. To measure these approaches with same criterion, lots of activity datasets are collected, forming public and transparent benchmarks for comparing different approaches.

In addition to the activity classification approaches, another critical research area within the HAR scope, the human tracking approach, is also reviewed briefly in a separate section. It is widely concerned especially in video surveillance systems for suspicious behavior detection.

The writing of rest parts conforms to general HAR process flow. First, research emphases and challenges of this domain are briefly illustrated in Section 2 . Then, effective features need to be designed for the representation of activity images or videos. Thus, Sections 3 and 4 , respectively, review the global and local representations in conventional RGB videos. Depth image-based representations are discussed as a separate part in Section 5 . Next, Section 6 describes the classification approaches. To measure and compare different approaches, benchmark datasets act an important role on which various approaches are evaluated. Section 7 collects recent human tracking methods of two dominant categories. In Section 8 we present representative datasets in different levels. Before we conclude this review and the future of HAR in Section 8 , we classify existing literatures with a detailed taxonomy ( Table 1 ) including representation and classification methods, as well as the used datasets aiming at a comprehensive and convenient overview for HAR researchers.

Feature encoding methods.

2. Challenges of the Domain

2.1. intraclass variation and interclass similarity.

Different from speech recognition, there is no grammar and strict definition for human activities. This causes twofold confusions. On one hand, the same activity may vary from subject to subject, which leads to the intraclass variations. The performing speed and strength also increase the interclass gaps. On the other hand, different activities may express similar shapes (e.g., using a laptop and reading). This is termed as interclass similarity which is a common phenomenon in HAR. Accurate and distinctive features need to be designed and extracted from activity videos to deal with these problems.

2.2. Recognition under Real-World Settings

2.2.1. complex and various backgrounds.

While applications like video surveillance and fall detection system use static cameras, more scenarios adopt dynamic recording devices. Sports event broadcast is a typical case of dynamic recording. In fact, with the popularity of smart devices such as smart glasses and smartphones, people tend to record videos with embedded cameras from wearable devices anytime. Most of these real-world videos have complex dynamic backgrounds. First, those videos, as well as the broadcasts, are recorded in various and changing backgrounds. Second, realistic videos abound with occlusions, illumination variance, and viewpoint changes, which make it harder to recognize activities in such complex and various conditions.

2.2.2. Multisubject Interactions and Group Activities

Earlier research concentrated on low-level human activities such as jumping, running, and waving hands. One typical characteristic of these activities is having a single subject without any human-human or human-object interactions. However, in the real world, people tend to perform interactive activities with one or more persons and objects. An American football game is a good example of interaction and group activity where multiple players (i.e., human-human interaction) in a team protect the football (i.e., human-object interaction) jointly and compete with players in the other team. It is a challenging task to locate and track multiple subjects synchronously or recognize the whole human group activities as “playing football” instead of “running.”

2.2.3. Long-Distance and Low-Quality Videos

Long-distance and low-quality videos with severe occlusions exist in many scenarios of video surveillance. Large and crowded places like the metro and passenger terminal of the airport are representative occasions where occlusions happen frequently. Besides, surveillance cameras installed in high places cannot provide high-quality videos like present datasets in which the target person is clear and obvious. Though we do not expect to track everyone in these cases, some abnormal or crime-related behaviors should be recognized by the HAR system ( Figure 2(b) ). Another typical long-distance case is the football broadcast ( Figure 2(a) ). Due to the long distance of cameras, the subject is rather small which makes it difficult to analyze activities of the torso [ 10 ], and the relatively low quality of those long distance videos further increases the difficulty.

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.002.jpg

Long-distance videos under real-world settings. (a) HAR in long-distance broadcasts. (b) Abnormal behaviors in surveillance.

3. Global Representations

Global representations extract global descriptors directly from original videos or images and encode them as a whole feature. In this representation, the human subject is localized and isolated using background subtraction methods forming the silhouettes or shapes (i.e., region of interest (ROI)). Some global approaches encode ROI from which they derive corners, edges, or optical flow as descriptors. Other silhouette-based global representation methods stack the silhouette image along the time axis to form the 3D space-time volumes, then the volumes are utilized for representation. Besides, discrete Fourier transform (DFT) takes advantage of frequency domain information of ROI for recognition, also being a global approach. Global representation approaches were mostly proposed in earlier works and gradually outdated due to the sensitiveness to noise, occlusions, and viewpoint changes.

3.1. 2D Silhouettes and Shapes

To recognize the human activities in videos, an intuitive idea is to isolate the human body from the background. This procedure is called background subtraction or foreground extraction. The extracted foreground in the HAR is called silhouette, which is the region of interest and represented as a whole object in the global representation approach.

Calculating the background model is an important step before extracting silhouettes. Wren et al. [ 11 ] first proposed to model the background scene with Gaussian distribution. Koller et al. [ 12 ] pointed out that some foreground values update unduly and thus they introduced the selective background update strategy. Stauffer and Grimson [ 13 ] proposed to model the values of a particular background pixel as a mixture of Gaussians to replace the strategy of using only one Gaussian value in the previous approach. The Gaussian mixture model (GMM) has been applied widely but the introduction of expectation maximization (EM) algorithm increases the computational cost. To reduce the cost, k -means clustering algorithm is used to replace the EM algorithm with an insignificant loss of accuracy. It is worth mentioning that current RGBD cameras make it easy to obtain the silhouette by using the depth data provided by depth sensors.

Besides the silhouette representation, the 2D shape of the silhouette can be used as a feature as well. Veeraraghavan et al. [ 14 ] emphasized the effectiveness of shape features. In their experiments, shape and kinematics that are being considered as two important cues in human motion were evaluated. Tests on both the gait-based human identification and the activity recognition indicate that shape plays a more important role. Veeraraghavan et al. then used this shape representation in their following work [ 15 ].

Bobick and Davis [ 16 , 17 ] stacked the silhouettes as two components for recognizing activities, respectively, the motion-energy image (MEI) and the motion-history image (MHI), which are both 2D representations.

In [ 18 ], oriented rectangular patches are extracted over the silhouettes. Spatial oriented histograms are then formed to represent the distribution of these rectangular patches. Those descriptors are finally used to recognize activities.

Extracting silhouettes from a single view is hard to satisfy view invariant property. To alleviate the influence of viewpoint changes, multiple cameras can be used to extract silhouettes in different viewpoints. Xu and Huang [ 19 ] proposed an “envelop shape” representation using two orthogonally placed cameras, which is robust to view changes of yaw rotation. Weinland et al. [ 20 ] made the same assumption that only the variations in viewpoints around the central vertical axis of the human body need to be considered. Motion history volumes (MHVs) were derived by stacking 4D silhouettes from four orthogonal cameras. In [ 21 ], a data fusion method was proposed, calculating the minimum DTW score between the test template and the two orthogonal view training templates.

3.2. Optical Flow

Optical flow is an effective way to extract and describe silhouettes for a dynamic background. Lucas-Kanade-Tomasi (LKT) feature tracker [ 22 , 23 ] can be used to obtain the optical flow. Lu et al. [ 24 ] used a LKT feature tracker approach to track joints in key frames and actual frames. Each activity is represented as a posture sequence, and each key posture is recorded in a key frame. Specific posture in actual frames can be recognized by finding correspondence between the actual and key frame. The recognized posture from the actual frame is compared to the key posture frame by mapping body locations, and the matched posture sequences are confirmed as the activity.

For recognizing human activities at a distance (i.e., the football broadcast video), Efros et al. [ 10 ] introduced a descriptor based on computing the optical flow to describe the “small” football players in person-centered images. Obviously, the background is dynamic due to the movement of players which makes it hard to model for background subtraction.

Tran and Sorokin [ 25 ] combined silhouettes and optical flow features together. Normalized bounding box is scaled to capture the region of the human body, and the optical flow measurements within the box are split into horizontal and vertical channels, while the silhouette gives the third channel. Subwindows are further divided to calculate histograms, and concatenating histograms of all 3 channels form the final descriptor.

3.3. 3D Space-Time Volumes (STVs)

An activity video can be seen as a series of images that contain activity sequences. Concatenating all frames along the time axis forms the 3D space-time volume (STV) which has three dimensions including two spatial dimensions X and Y and one temporal dimension T . Representations based on STVs expect to capture the additional dynamic information which the spatial representation methods cannot obtain due to the absence of time dimension. Constructing STVs for different activities is a global representation method. However, the STV sometimes combines with local features to build the final feature sets.

Blank et al. [ 5 ] first introduced the space-time shape to represent human activities. Space-time shape is obtained by only stacking the silhouette regions within images. However, due to the nonrigidity of the constructed 3D space-time shapes and inherent difference between space and time dimensions, traditional 3D shape analysis cannot be applied to the space-time activity shapes. Thus, the solution of the Poisson equation is used to derive local space-time saliency and orientation features.

Achard et al. [ 26 ] generated semiglobal features named space-time micro volumes from image sequence to deal with performances of different temporal durations. Motivated by seeking the common underlying induced motion fields of sequences of the same behaviors, Shechtman et al. [ 27 ] proposed an approach to compare volumes according to their patches. This method requires no prior modeling or learning of activities, being able to handle the complex dynamic scenes and detect multiple activities that occur simultaneously within the camera view. Their method is partially invariant to the changes in scale and orientation.

In [ 28 ], the input videos are segmented into space-time volumes using mean shift clustering technique. These oversegmented regions, which are termed “super-voxels,” are then matched using a proposed shape-matching technique, which is compared to the traditional silhouette matching methods. Unlike the previous silhouette-based approaches, the proposed shape-based representation does not require background subtraction nor explicit background models. To avoid the shortages of the shape-matching methods that are ignoring features inside the shape, Shechtman and Irani's flow-based features [ 27 ] are further incorporated.

3.4. Discrete Fourier Transform (DFT)

The DFT of image frame is another global feature that contains the intensity information of the foreground object (i.e., the region of the subject's body) provided that the foreground object intensity is different from the background. Kumari and Mitra [ 29 ] took advantage of this hypothesis and proposed a DFT-based approach, obtaining information about the geometric structure of the spatial domain foreground object. Normalized image frame is divided into small size blocks within which the average of all the DFT values is calculated. Finally the K-nearest neighbor (KNN) is applied to classify the DFT features and generate the activity classification result. The extracted DFT feature is novel compared to the previous work; however, its performance is restricted to simple backgrounds. The background in their test video datasets is almost blank.

4. Local Representations

Instead of extracting the silhouette or STV and encoding them as a whole, local representations process activity video as a collection of local descriptors. They focus on specific local patches which are determined by interest point detectors or densely sampling [ 30 ]. Most existing local features are proved to be robust against noise and partial occlusions comparing to global features. Local features are then normally combined with the bag-of-visual-words (BoVW) model and yield the general pipeline of current state-of-the-art local representation approaches [ 31 ]. Oriented from bag-of-words (BoW), BoVW-based local representation mainly contains four steps: feature extraction, codebook generation, feature encoding, and pooling and normalization. We follow [ 32 ] and state a traditional BoVW pipeline here: interest points and local patches are first obtained by detectors or densely sampled. Then local features are extracted from those interest points or patches. Next, a visual dictionary (i.e., codebook) is learned in training set by k -means or Gaussian mixture model (GMM), the original high-dimension descriptors are clustered, and the center of each cluster is regarded as a visual codeword. After that, local features are encoded and pooled. Finally, the pooled vectors are normalized as video representation. Among these steps, the development of more elaborately designed low-level features and more sophisticated encoding methods are the two chief reasons for the great achievements in this field [ 32 , 33 ], so in this part, we review the feature extraction methods in Section 4.1 and Section 4.2 , as well as the encoding methods in Section 4.3 .

4.1. Spatiotemporal Interest Point Detector

An intuitive thought of local representation is to identify those interest points that contain high information contents in images or videos. Harris and Stephens [ 34 ] first proposed effective 2D interest point detectors, the well-known Harris corner detector, which is extensively used in object detection. Then, Laptev and Lindeberg [ 6 ] proposed the 3D space-time interest points (STIPs) by extending Harris detectors. Spatial interest points in images are extended to spatiotemporal local structures in videos where the image values have significant local variations in both space and time. The spatiotemporal extents of the detected points are estimated by maximizing a normalized spatiotemporal Laplacian operator over spatial and temporal scales.

Saliency can also be used to detect interest points. Saliency means that certain parts of an image are preattentively distinctive and are immediately perceivable [ 35 ]. The spatiotemporal salient point can be regarded as an instance of the spatiotemporal interest point since both of them are informative and contain significant variations. The 2D salient point detection was first proposed by Kadir and Brady in [ 35 ]. Oikonomopoulos et al. [ 36 ] extended the 2D saliency to 3D spatiotemporal salient points that are salient both in space and time field. The salient points are successfully used as local features in their proposed activity classification scheme. Blank et al. [ 5 ] used the solution to Poisson equation to extract local space-time saliency of moving parts in the space-time shape. The detected salient points along with the local orientation and aspect ratios of shapes are calculated as local features.

Although these methods achieved remarkable results in HAR, one common deficiency is the inadequate number of stable interest points. In fact, the trade-off between the stability of those points and the number of points found is difficult to control. On one hand, the “right” and “discriminative” (i.e., stable) interest points are rare and difficult to be identified. As stated in [ 37 ], the direct 3D counterparts to commonly used 2D interest point detectors are inadequate, and true spatiotemporal corners are quite rare in certain applications. On the other hand, false alarms occur frequently due to various factors such as unintentional appearance changes. Ke et al. [ 38 ] illustrated two instances to point out that original detectors may fail in situations where the motions contain no sharp extrema; however, these detectors can be triggered falsely by the appearance of shadows and highlights in video sequences.

Besides the inherent properties of sparse interest points, many of the mentioned methods are inefficient. Therefore, these methods are restricted to the detection of a small number of points, or limited to low-resolution videos [ 39 ]. Here, we introduce some works either efficiency-enhanced or increasing number of stable interest points in response to the mentioned deficiency.

Dollar et al. [ 37 ] observed the rarity of the spatiotemporal interest points and the consequent problems of it in the recognition scheme. To find more 3D interest points in cuboids of space and time for activity recognition, the response function calculated by the separable linear filters is applied. The filtering is applied separately on the spatial and temporal dimensions, that is, 2D Gaussian smoothing kernel applied in spatial dimensions, and 1D Gabor filters applied in temporal dimension. Number of interest points increases using their detectors. Ke et al. [ 38 ] doubted the assumption that one can reliably detect a sufficient number of stable interest points in the video sequence. They extended the notion of rectangle features [ 40 ] into spatiotemporal volumetric features and applied the proposed framework on the video's optical flow. Their classifier is not limited to the sparseness nor affected by the instability of detected points.

Aiming at detecting interest points in an efficient way, Willems et al. [ 39 ] presented a dense, scale-invariant yet efficient spatiotemporal interest point detector with minimal effect on the computation time. First, point localization and scale selection are combined in a direct way using the determinant of the 3D Hessian matrix, therefore removing the time-consuming iterative scheme [ 41 ]. Further, building on Ke et al.'s work [ 38 ], an implementation scheme using integral video is developed to compute scale-invariant spatiotemporal features efficiently. Using a completely different idea, Oshin et al. [ 42 ] proposed to learn a classifier capable of detecting interest points in a novel video, given examples of the type of interest point that wish to get within a training video. The spatiotemporal Fern classifier (i.e., a seminaïve Bayesian classifier in [ 43 ]) is trained to recognize spatiotemporal interest points and thus achieves a high efficiency in constant time regardless of original detector complexity.

4.2. Local Descriptors

Local descriptors are designed to describe the patches that sampled either densely or at the interest points [ 1 ]. Effective descriptors are considered to be discriminative for the target human activity events in videos and robust to occlusion, rotation, and background noise.

Laptev [ 41 ] represented their 3D Harris corner by computing local, spatiotemporal N-jets as the descriptor. The descriptor is scale-invariant since they estimate the spatiotemporal extents of detected events by maximizing a normalized spatiotemporal Laplacian operator over spatial and temporal scales. Moreover, the proposed descriptors are proved to be robust to occlusions and dynamic cluttered backgrounds in the human motion analysis.

Similar to works of extending 2D interest point detector into spatiotemporal domain, such as the Harris corner detector [ 34 ] and the extended spatiotemporal one [ 41 ], many spatiotemporal descriptors were proposed by extending mutual image descriptors as well. We briefly review these works including both the original spatial descriptors and the spatiotemporal version of them.

Lowe proposed the scale-invariant feature transform (SIFT) in 1999 [ 44 ] and further improved it in 2004 [ 45 ]. It is widely used in local representation due to its scale and rotation invariance, as well as the robustness to affine distortion, changes in 3D viewpoint, addition of noise, and change in illumination. Scovanner et al. [ 46 ] introduced a 3D SIFT descriptor and used it in HAR. The 2D gradient magnitude and orientation are extended in 3D formulation; thus, creating the subhistograms encode the 3D SIFT descriptor. The videos are then described as a bag of spatiotemporal words using the 3D SIFT descriptor. Moreover, a feature grouping histogram which groups the co-occurred words out of the original one is used to build a more discriminative action video representation and finally used for classification.

The speeded-up robust features (SURF) [ 47 ] approach is a scale and rotation invariant detector and descriptor. The most important property of SURF is the improvement of efficiency comparing to previous approach. In the interest point detection, the approach applies the strategy that analyzing the input image at different scales to guarantee invariance to scale changes. Taking computation time into account, a very basic Hessian-matrix approximation which lends itself to the use of integral images is used for interest point detection, and it reduced the computation time dramatically. Next, a rotation and scale-invariant descriptor is provided for the detected interest point. The SURF approach builds on the distribution of first-order Haar-wavelet responses within the interest point neighborhood, in contrast with SIFT that extracts gradient information. Furthermore, integral images are exploited for speed. The introduction of indexing step based on the sign of the Laplacian further increases the robustness of descriptor and the matching speed.

An extended 3D SURF descriptor was implemented by Willems et al. [ 39 ]. Both of the 2D and 3D SURF used Haar-wavelet responses; however, the 3D SURF store the vector of the 3 axis responses instead of including the sums over the absolute values since the latter proved to be of no significant benefit but doubling the descriptor size.

Dalal and Triggs [ 48 ] proposed the histogram of oriented gradients (HOG) descriptor and achieved great success in human detection with linear SVM classifier. The good performance is due to the fact that the HOG's density distribution of local intensity gradients or edge directions can well characterize the local object appearance and shape of target objects.

Lu and Little et al. [ 49 ] presented the PCA-HOG descriptor which projects the original histogram of oriented gradients (HOG) descriptor to a linear subspace by principle component analysis (PCA). The descriptor was used to represent athletes to solve the problem of tracking and activity recognition simultaneously. Using HOG and HOF (histogram of flow) descriptor, Laptev et al. [ 50 ] completes a similar but more challenging activity recognition task as those activities are extracted from movies.

Klaser et al. [ 30 ] generalized the HOG descriptor to video sequences and proposed the HOG3D. Integral images are extended to integral videos for efficient 3D gradient computation. Polyhedrons are utilized for orientation quantization as an analogy of polygons in 2D space HOG. Optimized parameters for activity recognition have also been explored in their work.

Early spatiotemporal methods adopt a perspective of regarding the video as x - y - t 3D volumes [ 30 , 39 , 46 ]. However, recent feature trajectory approach considers the spatial dimensions x - y very different from the temporal dimension t . This approach detects the x - y interest points from video frames and then tracking them through video sequences as a trajectory. For detecting interest point, classic 2D detectors such as HOG and HOF are still used. In this review, we treat the feature trajectory as a special kind of the spatiotemporal descriptors where the time dimension is used to concatenate those 2D interest points.

Wang et al. [ 51 ] proposed dense trajectories by densely sampling points. Avoiding extracting points frame by frame and concatenating them, Wang et al. firstly extracted dense optical flow using Farneback's algorithm [ 52 ], then points can be densely tracked along the trajectory without additional cost. HOG and HOF are computed along the dense trajectories as the descriptors. Dense trajectories were further improved in [ 53 ]. The camera motion, as a main obstacle for extracting target trajectories from humans or objects of interests, was highlighted and was tried to be removed. The authors first match feature points using two complementary descriptors (i.e., SURF and dense optical flow), then estimate the homography using RANSC [ 54 ]. Through this approach, the camera motion is explicitly identified and removed. However, in some cases where humans dominate the frame, the target human motion may also generate inconsistent camera motion match. To solve this problem, a human detector is further explored to remove the inconsistent matches within the detected human areas. Improved descriptors achieved significant performance on challenge datasets, such as Hollywood2 where camera motions were used abundantly. Shi et al. [ 55 ] presented a sequential deep trajectory descriptor (sDTD) on the dense trajectory basis to capture the long-term motion information. The dense trajectories are projected into two-dimensional planes and a CNN-RNN network is employed to learn an effective representation for long-term motion.

4.3. Feature Encoding Methods

The STIP-based descriptors or other elaborately designed descriptors are all referred as local features. Local features are then encoded with feature encoding methods to represent activities and the encoded features are subsequently fed into pretrained classifiers (e.g., SVM) [ 32 ]. Encoding feature is a key step for constructing BoVW representation and utilizing an appropriate encoding method can significantly improve the recognition accuracy [ 56 ]. Here, we summarize the common feature encoding methods in recent literatures in Table 2 . The number of citations for each description paper is also provided to facilitate measurement of their influences.

Taxonomy of activity recognition literatures.

Several evaluations [ 56 – 58 ] have been conducted to compare the performance of recent encoding methods. Chatfield et al. [ 57 ] compared five encoding methods including LLC, SVC, FV, KCB, and the standard spatial histograms baseline. Experiments over PASCAL VOC 2007 and Caltech 101 show that FV performs best. Wang et al. [ 56 ] drew the same conclusion on KTH dataset and HMDB51 dataset. Also, a most recent evaluation [ 58 ] showed a consistent finding on UCF-YouTube and HMDB51 datasets, though slightly slower than local NBNN on KTH.

Further exploration has been conducted to match the best local feature with FV. In [ 31 ], six representative methods including VQ, SA- k , LLC, FV, VLAD, and SVC are evaluated for two widely used local features, STIPs and improved dense trajectories (iDTs). The experiment results demonstrate that the iDT together with the FV yields the best performance on the test datasets. Wang et al. who proposed the iDT also verified the best performance of iDT and FV in their work [ 53 ].

Recent stacked Fisher vectors [ 32 ] further improved the performance of iDT + FV and achieved superior performance when combining traditional FV. Evaluation on the YouTube, J-HMDB, and HMDB51 datasets demonstrates that it has become the state-of-the-art method. Pipelines of SFV and corresponding FV are given in Figure 3 .

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.003.jpg

Pipeline of Fisher vector and Stacked fisher vector. (a) Fisher vector. (b) Stacked fisher vector.

The core idea of both FV and SFV is trying to catch more statistical information from images; in contrast, BoVW only retains the zero order statistics. Take an l -dimension local descriptor as an example. Assuming that the size of prelearned GMM is K ( K is the size of codebook). For the conventional BoVW, the final encoded feature is K -dimension histograms that indicate the frequency of codewords. However, FV can obtain a 2 Kd -dimension ( d is the Gaussian distribution dimension). In another word, FV retained more information (i.e., high-order statistics) regarding to same size of codebooks.

SFV further improved FV owing to a simple and intuitive reason that SFV densely calculated local features by dividing and scanning multiscale subvolumes. The main challenge is the holistic combination of those local FVs since encoding them using another FV directly is impossible because of the high dimension of them (2 Kd -dimension). Thus, a max-margin method is tactfully used to reduce dimensionality. As the local FVs are more densely sampled than the conventional FV and consequently contain more high order statistics, therefore, iDT with SFV achieves even better result than the state-of-the-art iDT with FV.

5. Depth-Based Representations

Previous research of HAR mainly concentrates on the video sequences captured by traditional RGB cameras. Depth cameras, however, have been limited due to their high cost and complexity of operation [ 74 ]. Thanks to the development of low-cost depth sensors such as Microsoft Kinect [ 75 ], an affordable and easier way to access the depth maps is provided. Furthermore, Kinect SDK released the application that can directly obtain the skeletal joint positions in real-time (adopting algorithms in [ 75 ]). The available depth maps and the skeletal information (see Figure 4 ) vigorously contributed to the computer vision community. These two features and their derivative features also triggered a wide interest to solve HAR problems using depth-based solutions, replacing conventional RGB-based methods, or acting as supplements to enhance the RGB-based methods. In this section, we separately reviewed the recent advance of activity representations using depth maps or skeletons.

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.004.jpg

Kinect RGBD cameras and their color images, depth maps, skeletal information. (a) Kinect v1 (2011). (b) Kinect v2 (2014). (c) Color image. (d) Depth map. (e) Skeleton captured by Kinect v1. (f) Skeleton captured by Kinect v2.

5.1. Representations Based on Depth Maps

Depth maps contain additional depth coordinates comparing to conventional color images and are more informative. Approaches presented in this section regard depth maps as spatiotemporal signals and extract features directly from them. These features are either used independently or combined with RGB channel to form multimodal features.

Li et al. [ 76 ] employed the action graph model, which represents activities using several salient postures serving as nodes in action graph. All activities share same posture sets and each posture is characterized as a bag of 3D points from the depth maps. However, involving all the 3D points is computationally expensive; thus, a simple and effective method to sample the representative 3D points is proposed, achieving over 90% recognition accuracy by sampling approximately 1% points according to their report.

Zhao et al. [ 77 ] proposed a framework of combing RGB and depth map features for HAR and presented an optimal scheme. For the RGB channels, spatiotemporal interest points are generated solely from it and the HOG and HOF are calculated to form the RGB based descriptors. For the depth channel, they proposed a depth map-based descriptor called local depth pattern (LDP), which simply calculates the difference of average depth values between a pair of cells within the STIP surrounding region.

Yang et al. [ 78 ] proposed to use HOG on depth maps. Depth maps are projected onto three orthogonal planes and the depth motion maps (DMM) are generated by accumulating global activities through entire video sequences. HOG are then computed from DMM as the representation of an action video. Another depth image-based work similar to the HOG is [ 74 ] where the histogram of oriented 4D normals (HON4D) descriptor, as a further generalization of HOG3D to four-dimensional depth videos, is proposed. HON4D descriptor calculates the histograms of oriented 4D surface normals in 4D space of time, depth, and spatial coordinates. A quantization of the 4D space is also presented. The approach in [ 79 ] is also based on the polynormal which is a cluster of neighboring hypersurface normals from a local spatiotemporal depth volume. A designed scheme aggregates the low-level polynormals in each adaptive spatiotemporal cell. The concatenation of feature vectors extracted from all spatiotemporal cells forms the final representation of depth sequences.

Jalal et al. [ 80 ] considered multifeatures from depth videos, extracting 3D human silhouettes and spatiotemporal joints values for their compact and sufficient information for HAR task.

5.2. Skeleton-Based Representations

Skeletons and joint positions are features generated from depth maps. Kinect device is popular in this representation due to its convenience of obtaining skeleton and joints. Application in Kinect v1 SDK generates 20 joints, while the later version (Kinect v2) generates 25 joints, adding 5 joints around the hands and neck (see Figure 4 ). We reviewed recent papers on skeleton-based representations and summarize three aspects efforts on improving the performance of skeleton-based representation.

First, skeleton model has an inherent deficiency that it always suffers the noisy skeleton problem when dealing with occlusions (see Figure 5 ) [ 76 ]. Features from inaccurate skeletons and joints may completely be wrong. Current approaches often solve it by combining other features that robust to occlusion or alleviate occlusion problem by separating the whole skeleton into different body parts and handling them independently since not all body parts are occluded.

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.005.jpg

Noisy skeleton problem caused by self-conclusion.

Second, an intuitive fact can be observed that not all skeletal joints are involved in a particular activity, and only a few active joints are meaningful and informative for a certain activity [ 81 ]. Concentrating on these active joints and abandoning the other inactive parts will generate more discriminative and robust features and are beneficial to deal with intraclass variations [ 82 ].

Finally, as an extracted feature from depth maps itself, skeleton-based representation is often combined with original depth information to form more informative and robust representation [ 82 , 83 ].

Xia et al. [ 84 ] proposed a skeleton-based representation named HOJ3D, the spherical histograms of 3D locations of selected joints. After reprojected using LDA and clustered into vocabularies, the encoded features are fed to hidden Markov model (HMM) for classification. The HOJ3D is robust to view changes due to the design of the spherical coordinate system and robust skeleton estimation.

Yang and Tian [ 85 ] proposed a new type of feature named EigenJoints. 3D position differences of joints are employed to characterize three kinds of activity information including posture feature, motion feature, and offset feature. To reduce redundancy and noise, PCA is further employed and the efficient leading eigenvectors are selected. Finally, the constructed features were fed into the naïve-Bayes-nearest-neighbor (NBNN) [ 86 ] and obtained improved performance.

Wang et al. [ 82 ] indicated that using joint positions alone is insufficient to represent an action, especially for the case involving interaction with objects. Consequently, they proposed a depth-based feature called local occupancy pattern (LOP) to describe the occupancy of the neighborhood of each point, for example, the occupied space around the hand joint when lifting a cup. The local occupancy information is described by the 3D point cloud around a particular joint. Moreover, to select the active and discriminative joint feature subset (i.e., actionlet) for a particular activity, a data mining solution is leveraged and then actionlet ensemble which is linear combination of actionlets is obtained to represent each activity. Similar to actionlet, Zhu et al. [ 87 ] learned the co-occurrences of joints by designing regularization in deep LSTM (long short-term memory) RNNs (recurrent neural networks).

Shahroudy et al. [ 83 ] proposed a multimodal multipart approach for activity recognition in depth map sequences, which combines the complementary skeleton-based features LOP in [ 82 ] and depth-based features local HON4D in [ 74 ] of each part together and builds up a multimodal multipart combination. The multimodal multipart features are formulated into their framework via the proposed hierarchical mixed norm.

Chen et al. [ 81 ] proposed a skeleton-based two-level hierarchical framework. In the first layer, a part-based clustering feature vector is introduced to find out the most relevant joints and clustered them to form an initial classification. Note that the recognition task is divided into several smaller and simple tasks, which are performed within a specific cluster. It is of benefit to solving the high intraclass variance since distinct sequences of the same action are grouped into different clusters. In the second layer, only the relevant joints within specific clusters are utilized for feature extraction, which enhances the validity of the features and reduces the computational costs.

Besides depth-based features, skeleton data can be combined with other RGB features. To deal with the noisy skeleton problem, Chaaraoui et al. [ 88 ] proposed to combine skeletal and silhouette-based features using feature fusion methods. The noisy skeleton problem caused by occlusions of body part is partially elevated by the silhouette-based features. Shahroudy et al. [ 83 ] separately extracted dense trajectories features from RGB channel and 3D locations of skeleton joints from depth channel. A hierarchical feature fusion method based on structured sparsity was developed to fuse these two heterogeneous features.

6. Activity Classification Approaches

The next stage of HAR is the classification of activities that have been represented by proper feature sets extracted from images or videos. In this stage, classification algorithms give the activity label as final result. Generally speaking, most activity classification algorithms can be divided into three categories namely template-based approaches, generative models and discriminative models. Template-based approaches is a relatively simple and well accepted approach; however, it can be sometimes computationally expensive. Generative models learn a model of the joint probability P(X,Y) of the inputs X and the label Y , then P(Y|X) is calculated using Bayes rules and the algorithms finally picking the most likely label Y [ 89 ]. In contrast, discriminative models determine the result label directly. Typical algorithms of generative models are hidden Markov model (HMM) and dynamic Bayesian network (DBN), while support vector machine (SVM), relevance vector machine (RVM), and artificial neural network (ANN) are typical discriminative models.

6.1. Template-Based Approaches

Template-based approaches try to portray common appearance characteristics of a certain activity using various representations. These common appearance characteristics, such as 2D/3D static images/volumes or a sequence of view models, are termed as templates. Most template-based methods extract 2D/3D static templates and compare the similarity between the extracted images/volumes of test videos and the stored templates. For the classification based on a sequence of key frames, dynamic time warping (DTW) is an effective approach.

6.1.1. Template Matching

Bobick and Davis [ 16 , 17 ] proposed a temporal-template-based approach. Two components, the motion-energy image (MEI) which represents the presence of motion and the motion-history image (MHI) which indicates the recency of motion, are generated for each template of an activity. In fact, the generated template images can be regarded as weighted projection of the space-time shape.

Shechtman and Irani [ 27 , 90 ] constructed the 3D space–time intensity video volume template from a short training video clip. This small template is compared to every segment of same size in the test video over all three dimensions. The degree of similarity between two segments (i.e., the template and a same size video segment from the test video) is evaluated by the proposed intensity patch-based approach. It divides the segments into smaller patch units, then computes and integrates local consistency measures between those small space-time patches. This method has an impressive ability of detecting multiple different activities that occur at the same time.

Common template-based methods are unable to generate single template for each activity. They often suffer the high computational cost due to maintaining and comparing various templates. Rodriguez et al. [ 91 ] proposed to use the maximum average correlation height (MACH), which is capable of capturing intraclass variability by synthesizing a single action MACH filter for each activity class. They also generalized the MACH filter to video and vector valued data by embedding the spectral domain into a domain of Clifford algebras, building an effective approach in discriminating activities.

6.1.2. Dynamic Time Warping

Dynamic time warping (DTW) is a kind of dynamic programming algorithm for matching two sequences with variances. Rabiner and Juang [ 7 ] first developed it for speech recognition problem, representing the words as template sequence and assign matching scores for new word. DTW is also applicable to HAR problem since the human activities can be viewed as a sequence of key frames. The recognition problem is transformed to a template matching task.

Darrell and Pentland [ 92 ] proposed to build the representation of gestures using a set of learned view models. DTW algorithm is used to match the gesture template obtained from the means and variations of correlation scores between image frames and view models.

Veeraraghavan et al. [ 93 ] proposed the DTW-based nonparametric models for the gait pattern problem. They modified the DTW algorithm to include the nature of the non-Euclidean space in which the shape deformations take place. By comparing the DTW-based nonparametric and the parametric methods and applying them to the problem of gait and activity recognition, this work concluded that the DTW is more applicable than parametric modeling when there is very little domain knowledge.

Although the DTW algorithm needs a few amounts of training samples, the computational complexity increases significantly when dealing with growing activity types or those activities with high inter/intra variance, because extensive templates are needed to store those invariance.

6.2. Generative Models

6.2.1. hidden markov model approach.

The recognition task is a typical evaluation problem which is one of the three hidden Markov model problems and can be solved by the forward algorithm. HMMs were initially proposed to solve the speech recognition problem [ 8 ]. Yamato et al. [ 94 ] first applied the HMM to recognize activities. Features that indicate the number of pixels in each divided mesh are obtained as observations for each frame. Then, the HMMs are trained using the observation feature vector sequences for each activity, including the initial probability of hidden states, the confusion matrix, and the transition matrix. By applying the representation mentioned above, the HAR problem (recognition of various tennis strokes) is transformed into a typical HMM evaluation problem, which can be solved using standard algorithm.

A brief summary of the deficiencies of basic HMM and several efficient extensions are presented in [ 95 ]. The basic HMM is ill-suited for modeling multiple interacting agents or body parts since it is single variable state representation, as well as those actions that have inherent hierarchical structure. Take human interaction as an example, as a kind of complex activities, it always contains more than one person in the video, to which the basic HMM is ill-suited since the standard HMM is suitable for the time structure. Another deficiency is the exponentially decayed duration model for state occupancy. This duration model has no memory of the time that has already spent on the state, which is unrealistic for activities. This is implicitly obtained from the constant state transition probability and the first-order Markov assumption, which implies that the probability of a state being observed for a certain interval of time decays exponentially with the length of the interval [ 96 ].

Previous work has proposed several variants of HMM to handle the mentioned deficiencies [ 95 – 97 ]. Motivated by this human interaction recognition task that have structure both in time and space (i.e., modeling activities of two or more persons), Oliver et al. [ 97 ] proposed the coupled HMM (CHMM) to model the interactions. Two HMM models are constructed for two agents and probabilities between hidden states are specified.

Flexible duration models were suggested including the hidden semi-Markov model (HSMM) and the variable transition HMMs (VT-HMM). The hidden semi-Markov model (HSMM) is a candidate approach that has explicit duration model with specific distribution. Duong et al. [ 98 ] exploited both the inherent hierarchical structure and the explicit duration model and the switching hidden semi-Markov model (S-HSMM) is introduced with two layers to represent high-level activities and atomic activities separately. Another semi-Markov model (HSMM) based work is shown in [ 96 ].

Alternatively, Ramesh and Wilpon [ 99 ] broke the implicit duration model by specifying the dependency between the transition probability and the duration. The variable transition HMMs (VT-HMMs, originally called inhomogeneous HMM in [ 99 ]) was proposed and applied in speech recognition. In VT-HMM, the transition probability of two states depends on the duration which is no longer constant. Natarajan and Nevatia [ 95 ] then presented a hierarchical variable transition HMM (HVT-HMM) based on Pamesh and Wilpon's work to recognize two-hand gestures and articulated motion of the entire body. The HVT-HMM has three layers, including a composite event layer with a single HMM representing the composite actions, a primitive event layer using a VT-HMM to represent the primitive actions, and a pose track layer with a single HMM. The pose is represented using a 23 degrees body model, including 19 degrees for joint angles, 3 degrees for direction of translation ( x , y , z ), and 1 degree for scale.

6.2.2. Dynamic Bayesian Networks

A dynamic Bayesian network (DBN) is a Bayesian network with the same structure unrolled in the time axis [ 100 ]. An important extension of DBN is that its state space contains more than one random variables, in contrast with the HMM that has only one single random variable. Thus, the HMM can be viewed as a simplified DBN with constrained number of random variables and fixed graph structures.

Figure 6 presents a typical DBN. Suk et al. [ 101 ] proposed this structure for two hands gesture recognition, from which we can see that there are three hidden variables. The three hidden variables represent the motion of two hands and their spatial relation, while five features including two hands' motion and the position relative to the face, as well as the spatial relation between hands are designed as observations. Then, the DBN structure is built and simplified using the first-order Markov assumptions. They proposed the DBN tailored for hands gesture recognition in contrast with the previous fixed structure of CHMM [ 102 ] which is not deemed effective for other than tight-coupled two-party interactions.

An external file that holds a picture, illustration, etc.
Object name is JHE2017-3090343.006.jpg

A typical dynamic Bayesian network [ 101 ].

Park and Aggarwal [ 103 ] presented a hierarchical Bayesian network methodology for recognizing five two-person interactions. The proposed method first segments the body-part regions and estimates each of the body-part poses separately in the first level. Then, the individual Bayesian networks are integrated in a hierarchy to estimate the overall body poses of a person in each frame. Finally, the pose estimation results that include two-person interactions are concatenated to form a sequence with DBN algorithm.

Cherla et al. [ 21 ] indicated the contradiction for DTW between the robustness to intraclass variations and the computational complexity. Multiple templates for each activity handle the intraclass variations well but increase the computational complexity, while average templates reduce the complexity but are sensitive to intraclass variations. Cherla et al. proposed the average template with multiple feature representations to counterbalance them and achieve good performance.

6.3. Discriminative Models

6.3.1. support vector machines.

Support vector machines (SVMs) are typical classifiers of discriminative models and gained extensive use in HAR. Vapnik et al. [ 104 ] designed the SVM and originally used it for the problem of separating instances into two classes. It aims to find the hyperplane which maximizes the margin of two classes.

Schüldt et al. [ 105 ] combined SVM with their proposed local space-time features and applied their “local SVM approach” for HAR. A video dataset, known as the KTH dataset which had been one of the benchmarks of HAR systems, was recorded by them. The KTH dataset is introduced later in this paper (see Section 8.2.1 ).

Laptev et al. [ 50 ] used a nonlinear SVM with a multichannel Gaussian kernel and their SVM achieved high accuracy (91.8%) on the KTH dataset along with the HOG&HOF descriptors and local spatiotemporal bag-of-features. The well-known challenging Hollywood dataset (see Section 8.3.1 ) was provided and used to evaluate the proposed approach.

6.3.2. Conditional Random Fields

Conditional random fields (CRFs) are undirected graphical models that compactly represent the conditional probability of a particular label sequence Y , given a sequence of observations X . Vail et al. [ 106 ] compared the HMMs and CRFs for activity recognition. They found that the discriminatively trained CRF performed as well as or better than an HMM even when the model features are in accord with the independence assumptions of the HMM. This work pointed out a significant difference between the HMMs and CRFs: the HMMs assume that observations are independent given their labels; thus, complex features of the observation sequence will invalidate the assumption of this model and then make the HMM no longer a proper generative model. This inherent assumption of HMMs is abandoned in CRF, which conditions on the entire observation and therefore does not require any independence assumptions between the observation variables. A test was done by incorporating features which violate independence assumptions between observations (i.e., velocity thresholds in [ 106 ]) to explore the influence on both models. The result demonstrates that the CRF always outperforms the HMM, and with the increasingly severe violation of the independence assumptions, the HMM gets worse.

Natarajan and Nevatia [ 107 ] presented an approach for recognizing activities using CRF. Synthetic poses from multiple viewpoints are firstly rendered using Mocap data for known actions. Then, the poses are represented in a two-layer CRF, with observation potentials computed using shape similarity and transition potentials computed using optical flow. These basic potentials are enhanced with terms to represent spatial and temporal constraints, and the enhanced model is called the shape, flow, duration conditional random field (SFD-CRF). Single human activities as sitting down or standing up were recognized in their experiment.

Ning et al. [ 108 ] proposed a model that replaced the observation layer of a traditional random fields model with a latent pose estimator. The proposed model converted the high-dimensional observations into more compact and informative representations, and enabled transfer learning to utilize existing knowledge and data on image-to-pose relationship. This method has been shown to improve performance on the public available dataset HumanEva [ 109 ].

6.3.3. Deep Learning Architectures

Basically, the deep learning architectures can be categorized into four groups, namely deep neural networks (DNNs), convolutional neural networks (ConvNets or CNNs), recurrent neural networks (RNNs), and some emergent architectures [ 110 ].

The ConvNets is the most widely used one among the mentioned deep learning architectures. Krizhevsky et al. [ 9 ] first trained the deep ConvNets in a sufficiently large image datasets consisting of over 15 million labeled images. The impressive results lead to the extensively used of ConvNets in various pattern recognition domains [ 111 ]. Compared with traditional machine learning method and their hand-crafted features, the ConvNets can learn some representational features automatically [ 112 ]. Mo et al. [ 113 ] used ConvNets directly for feature extraction, and a multilayer perceptron is designed for the following classification.

One challenge for HAR using deep learning is how to apply it on small datasets since HAR datasets are generally smaller than what the ConvNets need. Common solutions include generating or dumpling more training instances, or converting HAR to a still image classification problem to leverage the large image dataset (e.g., ImageNet) to pretrain the ConvNets. Wang et al. [ 114 ] developed three strategies to leverage ConvNets on small training datasets. First, 3D points of depth maps are rotated to mimic different viewpoints, and WHDMMs at different temporal scales are constructed. Second, ConvNets model trained over ImageNet is adopted through transfer learning. Finally, different motion patterns are encoded into the pseudo-RGB channels with enhancement before being input to the ConvNets. On the other hand, Simonyan and Zisserm [ 115 ] leverage the large image dataset to pretrain the ConvNets. They investigated an architecture based on two separate streams (spatial and temporal), while the spatial stream contains information on appearance from still frames and is implemented using a spatial stream ConvNet. The spatial ConvNet is image classification architecture itself; thus, it is pretrained on the large image classification dataset.

The most recent research aims to further improve the performance of ConvNets by combining it with other hand-crafted features or representations. Li et al. [ 116 ] noted that the long-range dynamics information is necessary and should be modeled explicitly. Thus, they proposed a representation named VLAD 3 , which not only captures short-term dynamics with ConvNets but also utilizes the linear dynamic systems and VLAD descriptor for medium-range and long-range dynamics. Wang et al. [ 117 ] proposed a trajectory-pooled deep-convolutional descriptor (TDD) which combined the hand-crafted local features (e.g., STIP, improved trajectories) and deep-learned features (e.g., 3D ConvNets [ 76 , 118 ], two-stream ConvNets [ 115 ]). The proposed TDD integrates the advantages of these two features and adopts the state-of-the-art improved trajectories and two-stream ConvNets.

Unlike ConvNets, DNNs still use hand-crafted features instead of automatically learning features by deep networks from raw data. Berlin and John [ 119 ] used Harris corner-based interest points and histogram-based features as input. The proposed deep neural network with stacked auto encoders are used to recognize human-human interactions. Huang et al. [ 120 ] learned Lie group features (i.e., one of the skeletal data representations that are learned by manifold-based approaches) by incorporating a Lie group structure into a deep network architecture.

RNNs are designed for sequential information and have been explored successfully in speech recognition and natural language processing [ 121 , 122 ]. Activity itself is a kind of time-series data and it is a natural thought to use RNNs for activity recognition.

Among various RNNs architectures, the long short-term memory (LSTM) is the most popular one as it is able to maintain observations in memory for extended periods of time [ 123 ]. As an initial study for activity recognition, a LSTM network was utilized to classify activities in soccer videos [ 124 ]. Then, further research [ 123 ] explicitly demonstrated the robustness of LSTM even as experimental conditions deteriorate and indicated its potential for robust real-world recognition. Veeriah et al. [ 125 ] extended the LSTM to differential recurrent neural networks (RNNs). By computing the different orders of derivative of state which is sensitive to the spatiotemporal structure, the salient spatiotemporal representations of actions are learned, while in contrast, the conventional LSTM does not capture salient dynamic patterns of activity.

In addition to videos, RNNs can also be applied to skeleton data for activity recognition. Du et al. [ 126 ] proposed a hierarchical RNNs structure for skeleton-based recognition. The human skeleton from Kinect are divided into five parts and are fed into subnets separately. Representations from subnets are hierarchically fused into a higher layer and finally fed into a single-layer perceptron, whose temporally accumulated output is the final decision.

A detailed taxonomy about the representation, classification methods, and the used datasets of the introduced works in this review are presented in Table 1 .

7. Human Tracking Approaches

Besides the activity classification approaches, another critical research area is the human tracking approach, which is widely concerned in video surveillance systems for suspicious behavior detection. Human tracking is performed to locate a person along the video sequence over a time period, and then the resultant trajectories of people are further processed by expert surveillance systems for analyzing human behaviors and identifying potential unsafe or abnormal situations [ 127 ]. In this section, we briefly review recent literatures of two dominant approaches, namely kernel-based tracking and filtering-based tracking.

7.1. Filter-Based Tracking

Filtering is one of the widely used approaches for tracking, and the representative Kalman filter (KF) [ 128 ] and particle filter (PF) [ 129 ] are two commonly used classic filtering techniques.

KF is a state estimate method based on linear dynamical systems that are perturbed by Gaussian noise [ 130 ]. Patel and Thakore utilized traditional KF to track moving objects, in both the indoor and outdoor places. Vijay and Johnson [ 131 ] also utilized traditional KF for tracking moving objects such as car or human. However, the tested scenarios of these cases are relatively spacious and thus seldom occlusion occur. Despite the good results that are achieved by the KF-based method, it is strictly constrained with effective foreground segmentation, and its ability is limited when handling the occlusion cases. Arroyo et al. [ 127 ] combined Kalman filtering with a linear sum assignment problem (LSAP). To deal with the occlusion problem, visual appearance information is used with image descriptors of GCH (global color histogram), LBP (local binary pattern), and HOG (histogram of oriented gradients) representing the color, texture, and gradient information, respectively.

Particle filter, or sequential Monte Carlo method [ 132 ], is another typical filtering method for tracking. PF is a conditional density propagation method that is utilized to deal with non-Gaussian distributions and multimodality cases [ 130 ]. Ali et al. [ 133 ] combined a head detector and particle filter for tracking multiple people in high-density crowds. Zhou et al. [ 130 ] presented a spatiotemporal motion energy particle filter for human tracking, which fuses the local features of colour histograms as well as the spatiotemporal motion energy. The proposed particle filter-based tracker achieved robustness to illumination changes and temporal occlusions through using these features, as the motion energy contains the dynamic characteristics of the targeted human. As a specific branch of particle filter research, the sequential Monte Carlo implementation of the probability hypothesis density (PHD) filter, known as the particle PHD filter, is well developed for solving multiple human tracking problems. A series of research have been conducted by Feng et al. in [ 134 – 138 ].

7.2. Kernel-Based Tracking

Kernel-based tracking [ 139 ] or mean shift tracking [ 140 ] tracks the object (human) by computing the motion of one or more spatially weighted color histograms (i.e., single kernel/multiple kernels) from the current frame to next frame based on an iteratively mean-shift procedure. The kernel-based approach has fast convergence speed and low computation requirement inherited from the efficient mean shift procedure [ 141 ].

Traditional kernel-based tracking used symmetric constant kernel, and it tends to encounter problems of object scale and object orientation variation, as well as the object shape deformation. Research was conducted concerning these problems. Liu et al. [ 142 ] presented a kernel-based tracking algorithm based on eigenshape kernel. Yilmaz [ 143 ] introduced a kernel-based tracking algorithm based on asymmetric kernel for the first time. This kernel uses the initial region inside the outline of the target as kernel template and generates a precise tracking contour of the object. Yuan-ming et al. [ 144 ] noticed the shortage of the fixed asymmetric kernel. They combined the contour evolution technology with the mean shift and proposed an enhanced mean shift tracking algorithm based on evolutive asymmetric kernel. Liu et al. [ 145 ] presented an adaptive shape kernel-based mean shift tracker. Shape of the adaptive kernel is reconstructed from the low-dimensional shape space obtained by nonlinear manifold learning technique to the high-dimensional shape space, aiming to be adaptive to the object shape.

Early literatures reported tracking methods using single kernel scheme. However, the single kernel-based tracking could fail when the human is concluded, that is, the object could be lost or mismatch due to the partial observation. Thus, multiple-kernel tracking is adopted in most cases of recent researches. Lee et al. [ 146 ] evaluated two kernel and four kernel schemes [ 147 ] and presented a similar two and four kernal evaluation. Chu et al. [ 148 ] proposed to utilize projected gradient to facilitate multiple-kernel tracking in finding the best match under predefined constraints. The occlusion is managed by employing adaptive weights, that is, decreasing the importance of the kernel being occluded whilst enhancing the ones which are well-observed. Hou et al. [ 149 ] integrated the deformable part model (DPM) and designed multiple kernels, each of which corresponds to a part model of a DPM-detected human.

8. Representative Datasets in HAR

Public datasets could be used to compare different approaches in the same standards therefore accelerate the development of HAR methods. In this section, several representative datasets are reviewed, organized as a three-level category mentioned in the beginning of this review (i.e., action primitive level, action/activity level, and interaction level). There have been a published good survey [ 4 ] which presents the available important public datasets; however, it mainly focused on the conventional RGB-based datasets and missed current depth-based datasets. Thus, several important benchmark depth or RGB-D datasets are also reviewed in this section, with an overview of them ( Table 3 ).

Overview of representative datasets.

8.1. Action Primitive Level Datasets

While action primitives often act as components of high level human activities (e.g., the action primitives are served as a layer in hierarchical HMM to recognize activities [ 95 ] or interactions [ 97 ]), some typical and meaningful action primitives, such as poses and gestures [ 150 ], gait pattern [ 151 ], are studied as separate topics. These topics aroused wide research interest due to their importance in applications such as human-computer interaction and health care. Here, we present two recent gesture dataset based on RGB-D as the representative dataset in this level.

8.1.1. NTU-MSR Kinect Hand Gesture Dataset (2013)

The NTU-MSR Kinect hand gesture dataset [ 152 ] is considered as an action primitive level since it is developed for gesture recognition. Gestures in it were collected by Kinect, and each of them consists of a color image and the corresponding depth map. Totally, 1000 cases of 10 gestures were collected by 10 subjects, and each gesture was performed 10 times by a single subject in different poses. The dataset is claimed as a challenging real-life dataset due to their cluttered backgrounds. Besides, for each gesture, the subject poses with variations in hand orientation, scale, articulation, and so forth.

8.1.2. MSRC-Kinect Gesture Dataset (2012)

The MSRC-Kinect gesture dataset [ 153 ] is another typical action primitive level dataset, in which large amounts of limb level movements (e.g., karate kicking forwards with right leg) were recorded. There are totally 6244 instances of 12 gestures performed by 30 people, collected by Kinect. Positions of 20 tracked joints are provided as well.

8.2. Action/Activity Level Datasets

According to our definition, action/activity is middle level human activity without any human-human or human-object interactions. We first review two classic datasets, namely KTH human activity dataset and Weizmann human activity dataset. Though these two datasets have gradually faded out of state-of-the-art and are considered as easy tasks (e.g., 100% accuracy for Weizmann in [ 18 , 25 , 95 ]), they did play important roles in the history and act as benchmarks in earlier HAR works. Then, the well-known benchmark dataset for depth-based approaches, MSR Action3D dataset, is introduced next.

8.2.1. KTH Activity Dataset (2004)

The KTH dataset [ 105 ] is one of the most frequently cited datasets. It contains 6 activities (walking, jogging, running, boxing, hand waving, and hand clapping) performed by 25 subjects in controlled sceneries including outdoors, outdoors with scale variation, outdoors with different clothes, and indoors. One important factor in their success is the high intraclass variation in it which is one of the criteria for evaluation algorithms. Although the videos were still taken using static cameras, the high variation details, such as various scenarios and actors' clothes, as well as the different viewpoints, make itself a fair and convincing datasets for comparison. Most of the collected human activities in it were performed by a single person without any human-object interaction; thus, it is categorized in the activity/action level.

8.2.2. Weizmann Activity Dataset (2005)

The Weizmann activity dataset [ 5 ] was created by the Weizmann Institute of Science (Israel) in 2005. The Weizmann dataset consists of 10 natural actions (running, walking, skipping, bending, jumping-jack, galloping-sideways, jumping-forward-on-two-legs, jumping-in-place-on-two-legs, waving-two-hands, and waving-one-hand) with 10 subjects. Totally, 90 video sequences in a low resolution of 180 ∗ 144, 50 fps were recorded using a fixed camera and a simple background. To address the robustness of the proposed algorithm in [ 5 ], ten additional video sequences of people walking in various complicated scenarios in front of different nonuniform backgrounds were collected. Similar to the KTH dataset, most human activities in Weizmann were performed by a single person without any human-object interaction; thus, it is categorized in the activity/action level.

8.2.3. MSR Action3D Dataset (2010)

The MSR Action3D dataset [ 76 ] is widely used as the benchmark for depth-based HAR approaches. Depth maps of 20 activity classes performed by 10 subjects are provided in it (high arm waving, horizontal arm waving, hammering, hand catching, forward punching, high throwing, drawing cross, drawing tick, drawing circle, clapping hand, waving two hand, side-boxing, bending, forward kicking, side kicking, jogging, tennis swing, tennis serve, golf swing, pickup, and throw). MSR Action3D is a pure depth datasets without any color images in it.

8.3. Interaction Level Datasets

Interaction level datasets are relatively difficult tasks. Due to the human or human-object interactions, interaction level human activities are more realistic and abound in various scenarios such as sport events [ 91 ], video surveillance, and different movie scenes [ 50 ]. In this section, we review two conventional RGB datasets (i.e., Hollywood human activity dataset and UCF sports human activity dataset) and a RGB-D dataset (i.e., MSR DailyActivity3D dataset). Designed to cover indoor daily activities, MSR DailyActivity3D dataset [ 160 ] is more challenging and involves more human-object interactions compared to MSR Action3D [ 82 ].

8.3.1. Hollywood Human Activity Dataset (2008 and 2009)

Another well-known interaction level dataset is the Hollywood human activity dataset [ 50 , 158 ]. As a representative of realistic activity dataset, the Hollywood dataset is introduced here as a challenging task compared to previous datasets due to its frequently moved camera viewpoints, occlusions, and dynamic backgrounds with seldom provided information [ 1 ]. The initial version published in 2008 [ 50 ] contains approximately 663 video samples (233 samples in automatic training set, 219 samples in clean training set, and 211 samples in test set) of eight actions (answering phone, getting out of car, hugging, handshaking, kissing, sitting down, sitting up, and standing up) from 32 movies. Recognition of natural human activities in diverse and realistic video settings, which can be tested on this dataset, was discussed in [ 50 ]. Then, the extended Hollywood dataset was created in 2009 [ 158 ], involving four additional activities (driving a car, eating, fighting, and running) and more samples for each class, totally, 3669 video clips from 69 movies. Both human interaction (e.g., kissing, fighting) and human-object interactions (e.g., answering phone, driving a car) are included. Marszalek et al. [ 158 ] exploited the relationship between context of natural dynamic scenes and human activities in video based on this extended Hollywood dataset.

8.3.2. UCF Sports Dataset (2007)

The UCF sports dataset [ 91 ] is a specific interaction level dataset focused on various sports activities from television broadcasts. It is one of the datasets collected by Computer Vision Lab, University of Central Florida. There are over 200 video sequences in this dataset, covering 9 sport activities including diving, golf swinging, kicking, lifting, horseback riding, running, skating, swinging a basketball bat, and pole vaulting. While it covers only 9 human activities in sports scenes, it is still a challenging task for recognition due to its unconstrained environment and abound intraclass variability.

8.3.3. MSR DailyAction3D Dataset (2012)

The MSR DailyActivity3D dataset [ 160 ] is an interactive level dataset captured by Kinect device. In contrast with the previous MSR Action3D, this dataset provides three types of data including depth maps, skeleton joint positions, and RGB video. 16 activity classes performed by 10 subjects (drinking, eating, reading book, calling cellphone, writing on a paper, using laptop, using vacuum cleaner, cheering up, sitting still, tossing paper, playing game, lying down on sofa, walking, playing guitar, standing up, and sitting down) are recorded in it.

9. Conclusions and Future Direction

Human activity recognition remains to be an important problem in computer vision. HAR is the basis for many applications such as video surveillance, health care, and human-computer interaction. Methodologies and technologies have made tremendous development in the past decades and have kept developing up to date. However, challenges still exist when facing realistic sceneries, in addition to the inherent intraclass variation and interclass similarity problem.

In this review, we divided human activities into three levels including action primitives, actions/activities, and interactions. We have summarized the classic and representative approaches to activity representation and classification, as well as some benchmark datasets in different levels. For representation approaches, we roughly sorted out the research trajectory from global representations to local representations and recent depth-based representations. The literatures were reviewed in this order. State-of-the-art approaches, especially those depth-based representations, were discussed, aiming to cover the recent development in HAR domain. As the next step, classification methods play important roles and prompt the advance of HAR. We categorized classification approaches into template-matching methods, discriminative models, and generative models. Totally, 7 types of method from the classic DTW to the newest deep learning were summarized. For human tracking approaches, two categories are considered namely filter-based and kernel-based human tracking. Finally, 7 datasets were introduced, covering different levels from primitive level to interaction level, ranging from classic datasets to recent benchmark for depth-based methods.

Though recent HAR approaches have achieved great success up to now, applying current HAR approaches in real-world systems or applications is still nontrivial. Three future directions are recommended to be considered and further explored.

First, current well-performed approaches are mostly hard to be implemented in real time or applied to wearable devices, as they are subject to constrained computing power. It is difficult for computational constrained systems to achieve comparable performances of those offline approaches. Existing work utilized additional inertial sensors to assist in recognizing, or developed microchips, for embedded devices. Besides these hardware-oriented solutions, from a computer vision perspective, more efficient descriptor extracting methods and classification approaches are expected to train recognition models fast, even in real time. Another possible way is to degrade quality of input image and strike a balance among input information, algorithm efficiency, and recognizing rate. For example, utilizing depth maps as inputs and abandoning color information are ways of degrading quality.

Second, many of the recognition tasks are solved case by case, for both the benchmark datasets and the recognition methods. The future direction of research is obviously encouraged to unite various datasets as a large, complex, and complete one. Though every dataset may act as benchmark in its specific domain, uniting all of them triggers more effective and general algorithms which are more close to real-world occasions. For example, recent deep learning is reported to perform better in a four-dataset-combined larger datasets [ 114 ]. Another promising direction is to explore an evaluation criterion which enables comparisons among wide variety of recognition methods. Specifically, several vital measuring indexes are defined and weighted according to specific task, evaluating methods by measuring indexes such as recognition rate, efficiency, robustness, number, and level of recognizable activities.

Third, mainstream recognition system remains in a relatively low level comparing with those higher level behaviors. Ideally, the system should be able to tell the behavior “having a meeting” rather than lots of people sitting and talking, or even more difficult, concluding that a person hurried to catch a bus rather than just recognizing “running.” Activities are analogous to the words consisting behavior languages. Analyzing logical and semantic relations between behaviors and activities is an important aspect, which can be learned by transferring from Natural language processing (NLP) techniques. Another conceivable direction is to derive additional features from contextual information. Though this direction has been largely exploited, current approaches usually introduce all the possible contextual variables without screening. This practice not only reduces the efficiency but also affects the accuracy. Thus, dynamically and reasonably choosing contextual information is a future good topic to be discussed.

Finally, though recent deep learning approaches achieve remarkable performance, a conjoint ConvNets + LSTM architecture is expected for activity video analysis in the future. On the one hand, ConvNets are spatial extension of conventional neural networks and exhibit its advantage in the image classification tasks. This structure captures the spatial correlation characteristics, however, ignores the temporal dependencies of the interframe content for activity dynamics modeling. On the other hand, LSTM as a representative kind of RNN, is able to model the temporal or sequence information, which makes up the temporal shortage of ConvNets. LSTM is currently used in accelerometer-based recognition, skeleton-based activity recognition, or one-dimensional signal processing, but has not been widely concerned in combination with ConvNets for two-dimensional video activity recognition, which we believe is a promising direction in the future.

Acknowledgments

This research is supported by the National Natural Science Foundation of China (no. 61602430, no. 61672475, and no. 61402428); major projects of Shandong Province (no. 2015ZDZX05002); Qingdao Science and Technology Development Plan (no. 16-5-1-13-jch); and The Aoshan Innovation Project in Science and Technology of Qingdao National Laboratory for Marine Science and Technology (no. 2016ASKJ07).

Conflicts of Interest

The authors declare that there is no conflict of interest regarding the publication of this paper.

shape based activity recognition

Journal of Materials Chemistry C

A selectively bimodal flexible sensor based on il/swcnts/pedot:pss nanocomposites for materials and shape recognition †.

ORCID logo

* Corresponding authors

a i-Lab Suzhou Institute of Nano-Tech and Nano-Bionics (SINANO), Chinese Academy of Sciences (CAS), 398 Ruoshui Road, Suzhou 215123, P. R. China E-mail: [email protected] , [email protected]

b School of Nano-Tech and Nano-Bionics, University of Science and Technology of China (USTC), Hefei, Anhui 230026, P. R. China

c Jiangxi Institute of Nanotechnology, Xiaolan Economic and Technological Development Zone, 278 Luozhu Road, Nanchang 330200, China

d Gusu Laboratory for Materials Science, 388 Ruoshui Road, Suzhou 215123, P. R. China

e Institute for Functional Intelligent Materials, National University of Singapore, 21 Lower Kent Ridge Road, Singapore 119077, Singapore

Flexible tactile sensors have attracted much attention since they can imitate the sensory functions of the human hand skin, showing conceivable applications in humanoid robots, human–machine interactions and so on. Unfortunately, the practical performance of these sensors is restricted by the drawback of inconvenient-multifunctional responses, such as force and temperature-derived bimodal signals with high coupling that need to be decoupled via complex algorithms. Herein, in this study, a novel IL/SWCNTs/PEDOT:PSS nanocomposite film was fabricated and has a Seebeck coefficient of 29.1 μV K −1 and an electrical conductivity of 8448 S m −1 , in which PEDOT:PSS forms a stable flexible matrix, SWCNTs provide a continuous conductive network, and the incorporation of IL plays dual roles in promoting the dispersion uniformity of SWCNTs and the phase separation of PEDOT:PSS. Furthermore, via the piezoresistive mechanism and thermoelectric output principle, this nanocomposite film derived dual-modal flexible tactile sensor was endowed with the comprehensive features of pressure and temperature response behaviors without any mutual coupling manner, manifesting the skin-like functions of selective sensing capability. Finally, this flexible device is integrated into a humanoid hand for accurately distinguishing the materials and shapes of objects, demonstrating a recognition accuracy of over 98% via the aid of a machine-learning strategy. It can be believed that this study will inspire the development of next-generation biomimetic robots with tactile perception.

Graphical abstract: A selectively bimodal flexible sensor based on IL/SWCNTs/PEDOT:PSS nanocomposites for materials and shape recognition

Supplementary files

  • Supplementary information PDF (889K)

Article information

Download citation, permissions.

shape based activity recognition

A selectively bimodal flexible sensor based on IL/SWCNTs/PEDOT:PSS nanocomposites for materials and shape recognition

S. Yuan, Y. Tian, Y. Li, S. Li, L. Fu, T. Li and T. Zhang, J. Mater. Chem. C , 2024, Advance Article , DOI: 10.1039/D3TC04104B

To request permission to reproduce material from this article, please go to the Copyright Clearance Center request page .

If you are an author contributing to an RSC publication, you do not need to request permission provided correct acknowledgement is given.

If you are the author of this article, you do not need to request permission to reproduce figures and diagrams provided correct acknowledgement is given. If you want to reproduce the whole article in a third-party publication (excluding your thesis/dissertation for which permission is not required) please go to the Copyright Clearance Center request page .

Read more about how to correctly acknowledge RSC content .

Social activity

Search articles by author.

This article has not yet been cited.

Advertisements

IMAGES

  1. How to Teach Shape Recognition to Preschoolers with Fun Activities

    shape based activity recognition

  2. Shape Recognition Activity

    shape based activity recognition

  3. shapes worksheets Shape recognition games for preschoolers are very

    shape based activity recognition

  4. Shape Recognition Activity

    shape based activity recognition

  5. SHAPE RECOGNITION

    shape based activity recognition

  6. Shape Matching Activity. Free Printable to help toddlers with shape

    shape based activity recognition

VIDEO

  1. Enrgetic shape recognition by understanding

  2. Shape Recognition Activity For Toddlers / Easy Fine Motor Activity #learningathome #kindergarten

  3. shapes activity #learningmethod #youtubeshorts

  4. Colour Recognition Activity # simple activity for children #

  5. #class room shape recognition activitiy#shortvideo#youtube trending shorts#

  6. Activity Name -: colour recognition +Shapes RECOGNITION(cut and paste )

COMMENTS

  1. 10 Hands-On Shapes Recognition Activities for Preschoolers

    10 Hands-On Shapes Recognition Activities for Preschoolers By: Tanja McIlroy Last updated: 25 January 2023 Early Mathematical Skills Teaching children about shapes should be a fun, hands-on experience. It is by experiencing concepts through the body and the senses that real learning happens in the early years.

  2. A Week's Worth of Shape Recognition Activities

    Art Use shape toys for a fun art activity - super easy to set up! Supplies: Shape sorting toys, paint, paper Stamp paper tubes into paint for a simple process art circle activity. Supplies: Paper tubes, paint, paper Be creative and make a shape house out of construction paper. (My Mommy Style)

  3. 25 Shape Activities for Kids

    4. Kandinsky Shape Art for Kids - Messy Little Monster Explore colors and shapes with your toddler or preschooler with this shape art for kids. This hands on learning art project inspired by Kandinsky. Is fantastic for kids of all ages. 5. Recycled Shapes Process Art - Mosswood Connections Raid the recycle bin and use recycled items to make art.

  4. How to Teach Shape Recognition to Preschoolers with Fun Activities

    More shapes activities: Toddler Triangle Shapes Activity. 15 Activities for Teaching Squares. Shapes Art Using Toys. Teaching Circles to Toddlers and Preschoolers. Going on a Circle Hunt Circle Time Activity. Here's a simple shape matching activity for your children to try today! Click on the photo for the link to the PDF download:

  5. 25 Creative Activities and Ideas For Learning Shapes

    1. Start with an anchor chart Colorful anchor charts like these are terrific reference tools for kids learning shapes. Have kids help you come up with examples for each one. Learn more: A Spoonful of Learning / Kindergarten Kindergarten 2. Sort items by shape Collect items from around the classroom or house, then sort them by their shapes.

  6. 8 of the best shape recognition resources and activities for early

    8 of the best shape recognition resources and activities for early years Don't be a square, boost your kids' shape recognition skills with these ace activities by Teachwire DOWNLOAD A FREE RESOURCE! 2D shape recognition posters for Early Years and KS1 maths Download Now Download Now Primary Maths 1 | We're going on a shape hunt

  7. 25 Shape Activities For Preschool and Kindergarten

    Shapes are a fundamental aspect of geometry. Use shape recognition games, puzzles, and activities to introduce basic geometric concepts. Have children sort objects by shape, creating opportunities for classification and mathematical thinking. Manipulating and combining shapes helps develop spatial awareness and understanding, which is essential ...

  8. 15 Best Shape Activities For Preschoolers in 2024

    Instructions: Read a story with shape-themed books, pointing out and discussing the shapes you encounter. 13. Build with Blocks. Build with Blocks is a hands-on, interactive, and simple way to engage in shape activities for preschoolers. It fosters creativity, shape recognition, and bonding through constructive play.

  9. Hands-On Shape Activities for Preschool

    Play Shape Bingo: Call out a shape name (or describe the attributes only) and have students find the shape on their bingo cards. Shape Hunt: Attach paper shapes to the bottom of a small box or activity tub, cover with salt, then have students use a paintbrush to hunt for shapes, naming each shape as they find it.; Mystery Shapes: Draw a variety of shapes on white paper using white crayon.

  10. 17 Creative Shape Activities for Preschool and Kindergarten

    This Foam Lacing Shapes activity is great fine motor practice and makes a great busy bag. This Build a Bracelet activity is great for shape recognition and fine tuning your child's fine motor skills. This inexpensive Cardboard Shapes activity easy toddler craft is a cheap way to teach your toddlers shapes. Learn to recognize shapes in ...

  11. 20 Shape Activities for Preschool

    Empower your child with skills to thrive. 24 low-prep activities and relatable conversation starters that equip kids for life. Order the easy-to-follow activity cards. Build the foundation for social-emotional learning. Notice a shift in self-confidence and self-regulation as you and your child implement what you practice.

  12. 12 Fun Preschool Shapes Activities to Try

    Playful shape activities encourage them to explore the attributes of shapes, compare and classify them, and understand their real-world applications. So let them learn shapes through play! Try these shape activities for preschoolers below: 9. Shape Sensory Bin. Sensory bins are amazing for preschool children.

  13. 10+ Hands-On Shape Activities for Preschoolers

    Shape Activity 1: Shape collages Having a selection of paper cut outs to create pictures from is a great way to begin to engage your child in discussion about shapes! Shape Activity 2: Lollipop sticks Shape Activity 3: DIY paint stamps These require a little bit of preparation.

  14. 5 Shape Recognition Activities to Boost Development

    5 Shape Recognition Activities to Boost Development Children start showing an interest in shapes around 2 or 3 years old, but you can teach your little one about them from birth through everyday interactions. Your child can improve their visual-spatial abilities and spatial awareness by recognizing shapes.

  15. 40 Easy And Fun Hands-On Shape Activities For Preschoolers

    Shapes Activities for Preschoolers One of the first math concepts that preschoolers learn is identifying shapes. They begin to distinguish among the different shapes and categorize items according to shape. They learn the names of shapes and their characteristics. They find shapes in everyday items.

  16. Shapes Activities For Preschoolers

    Make shapes collages using contact paper. This sticky shapes activity is a shapes sorting activity. The contact paper removes the need for glue and makes it easier for a toddler. This learning shapes and colors for toddler Printable activity looks as pretty as it is fun. My toddler had a blast placing the small stars in the big star.

  17. 12 Shape Activities for Toddlers! It's Hip to be Square!

    Math & 123s Toddlers Resources. Shapes. These 12 hands on shape activities for toddlers (2 year olds) will make it hip to be square! Toddlers will have fun learning shapes in a hands on way. We've been having a little bit of fun lately doing some hands on shape activities! George is on a shapes kick right now, anytime he asks for an activity ...

  18. 22 Shape Activities For Preschoolers (2024)

    1. Materials you need: post-it's (my very best friend), construction paper, painter's tape, and a marker. View Amazon's Price 2. Set-up: on post-it's draw a bunch of different shapes that you want to work on. On construction paper, draw a big version of each shape that you drew on the post-it notes.

  19. Shape-Based Human Activity Recognition Using Independent ...

    In this paper, a novel human activity recognition method is proposed which utilizes independent components of activity shape information from image sequences and Hidden Markov Model (HMM) for recognition.

  20. Independent shape component-based human activity recognition via Hidden

    Various human activities are represented by shape feature vectors from the sequence of activity shape images via ICA. Based on these features, each HMM is trained and activity recognition is achieved by the trained HMMs of different activities.

  21. Skeleton-Based Activity Recognition: Preprocessing and Approaches

    2.1 Introduction Skeleton-based action recognition system has captured significant attention because of its excellent ability to deal with large data-set by improving the execution speed. The skeleton structures the entire organization of the human body's external architecture by providing shape, support, and other indispensable parts of the body.

  22. 25 Creative Ways to Teach 2D Shapes in Kindergarten

    Create a list of shapes and ask your students to find objects in the classroom or school that match those shapes. A shape hunt is a fun way to promote active learning and encourage students to observe their surroundings of real-world things. 3. Shape Dominos. This fun game makes a great math center.

  23. A Review on Human Activity Recognition Using Vision-Based Method

    Human activity recognition (HAR) aims to recognize activities from a series of observations on the actions of subjects and the environmental conditions. The vision-based HAR research is the basis of many applications including video surveillance, health care, and human-computer interaction (HCI). This review highlights the advances of state-of ...

  24. A selectively bimodal flexible sensor based on IL/SWCNTs/PEDOT:PSS

    Finally, this flexible device is integrated into a humanoid hand for accurately distinguishing the materials and shapes of objects, demonstrating a recognition accuracy of over 98% via the aid of a machine-learning strategy. It can be believed that this study will inspire the development of next-generation biomimetic robots with tactile perception.