Social Share


Problem statement

There is no one around the world who doesn’t know of the animated comedy series, Tom and Jerry. Let’s admit it—all of us still love the iconic show and wish to catch a glimpse of Tom’s and Jerry’s constant notorious banter. Jerry leaves no stone unturned to annoy Tom—be it getting Tom in trouble with his landlady Mammy Two Shoes and his arch-nemesis Spike, making a fool out of him in front of his love interest Toodles Galore, or beating him for bothering Nibbles or Quaker. No matter what, we always end up laughing till our stomachs hurt.

On this International Day of Happiness, we are bringing back all the joy and happiness with this challenge. In this challenge, you are required to build a model that detects emotions of the characters in a video frame from our most-loved show, Tom and Jerry. 

Your task is to extract frames from a video clip provided and classify the primary character’s emotion into one of the five classes: angry, happy, sad, surprised, or Unknown.


The dataset consists of two parameters—‘Frame_ID’ that indicates the frame of the video and ‘Emotion’ that categorizes the emotion of the primary character into different labels: angry, happy, sad, surprised, or Unknown.

The benefits of practicing this problem by using Machine Learning/Deep Learning techniques are as follows:

  • This challenge will encourage you to apply your Machine Learning skills to build models that analyze and classify frames of videos
  • This challenge will help you enhance your knowledge of multi-label image classification actively. It is one of the basic building blocks of Deep Learning

We challenge you to build a model that detects the emotions of Tom or Jerry in each video frame.


  • 1st Prize - 250 USD

  • 2nd Prize - 150 USD

  • 3rd Prize - 75 USD


Deep Learning is an application of Artificial Intelligence (AI) that provides systems with the ability to automatically learn and improve from experience without being explicitly programmed. Deep Learning is a science that determines patterns in data. These patterns provide deeper meaning to problems and help you to first understand problems better and then solve the same with elegance. HackerEarth’s Deep Learning challenge is designed to help you improve your Deep Learning skills by competing and learning from fellow participants.

Here’s presenting HackerEarth’s Deep Learning Challenge—Detect emotions of your favorite toons.

Why should you participate?

  • To analyze and implement multiple algorithms and determine which is more appropriate for a problem
  • To get hands-on experience in Deep Learning problems

Who should participate?

  • Working professionals
  • Deep Learning or Machine Learning enthusiasts
  • College students (if you understand the basics of predictive modeling)


Machine Learning practice


  • In order to claim your prizes, more than 50% of your HackerEarth profile must be complete
  • The prizes will be disbursed in the second week of the following month


  1. Your output will be evaluated only for 50% of the test data while the contest is running. Once the contest is over, output for the remaining 50% of the data will be eval...
View All Notifications