TJEDC - Tom and Jerry Emotion Detection Challenge

🕵️ Introduction

We all remember Tom and Jerry and the immense happiness their misadventures brought in our lives. Now you are required to build a model that detects emotions of the characters in a video frame from our most-loved show. Build a model that detects the emotion of the character present in each frame.

💾 Dataset

The dataset consists of two parameters—

• Frame_ID that indicates the frame of the video
• Emotion that categorizes the emotion of the primary character into different labels: angry(0), happy(1), sad(2), surprised(3), or Unknown(4). The number in brackets is the encoding of the emotion.

📁 Files

• train.csv - (1941 samples) File that should be used for training by the user.
• test.csv - (275 samples) File that will be used for actual evaluation for the leaderboard score.Note:- This is not the final test set. -train.zip - Contains the images corresponding to train.csv file -test.zip - Contains the images corresponding to test.csv file

🚀 Submission

• Prepare a csv containing header as emotion and predicted emotion as 0/1/2/3/4 with name as submission.csv.
• Sample submission format available at ./data/sample_submission.csv.

Make your first submission here 🚀 !!

🖊 Evaluation Criteria

During evaluation F1 score and Accuracy will be used to test the efficiency of the model

📱 Contact

• Pulkit Gera

📚 References

• https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html
• https://www.analyticsvidhya.com/blog/2019/10/building-image-classification-models-cnn-pytorch/