laxmikant-parate / facial-emotion-recognition Goto Github PK
View Code? Open in Web Editor NEWDetects human faces from input videos/images and predicts the Emotions. Used Sequential model from Keras, TensorFlow to build the model and trained it using the famous FER-2013 dataset with over28k images. Used Haar Cascades with cascade classifiers from OpenCV to locate face area in the input and integrated into web with HTML, CSS and Flask as backend. Improved overall emotion prediction accuracy to 94.23%.