Facial expression recognition using adapted residual based deep neural network

Emotion on our face can determine feelings, mental state and directly impact decisions. Humans are subjected to undergo an emotional change in relation their living environment or at a present circumstance. These emotions be anger, disgust, fear, sadness, happiness, surprise neutral. Due the intricacy nuance of facial expressions relationship emotions, accurate expression identification remains difficult undertaking. As result, we provide end-to-end system that uses residual blocks identify improve accuracy this research field. After receiving image, framework returns its state. The obtained test set FERGIT dataset (an extension FER2013 with 49300 images) was 75%. This proves efficiency model classifying as database poses bunch challenges such imbalanced data, intraclass variance, occlusion. To ensure performance model, also tested it CK+ output 97% set.