We have trained a Convolutional Neural Network to classify Facial Expressions.
Model Design:
Total params: 1,467,494
Trainable params: 1,467,494
Non-trainable params: 0
The types of expressions are:
- 'sad'
- 'happy'
- 'angry'
- 'surprised'
- 'fearful'
- 'neutral'
We have used the Facial-Expression-Recognition 2013 Dataset to train our model.
We have achieved an accuracy of 81.44%, 64.77% and 63.75% on train set, validation set and test set respectively.
Tech Stack:
- Tensorflow
- Python
- Streamlit
To use the Web Applization, visit the link below:
https://abhishekbiswas12-facial-expression-recognition-webapp-crlvvy.streamlit.app/