So in this project i have used OpenCV's Harcascade face classifier to detect the face in a live video through webcam, and then we do the inference on that image with a pre-trained model which is trained on FER 2013 dataset which i download from kaggle. So model does the inference and the predicted class is returned to the program which is the displayed on bounding boxes.
To download dataset:
Kaggle link: https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge/data
Train: https://drive.google.com/drive/folders/1F3ui2L5Z24mQmUjB7O43eMVChg5-e-rq?usp=sharing
Test: https://drive.google.com/drive/folders/1-sPkkNc17GoaUenhQDvK-llke_A1zZ46?usp=sharing