Identifying Human Facial Emotions by a Deep Neural Network

Abstract

We are moving towards an era where every aspect of our life is getting automated. So, getting facial emotion analysis done by artificial intelligence is one of the steps towards automation. Facial expressions are of the means of identifying one's emotion. Facial emotion recognition is widely used in fields like psychology, forensics and social media as it is of vital importance to get reliable results in all of these fields. The proposed system will work towards identifying the seven common emotions- happy, sad, anger, fear, disgust, surprise, neutral. Neural Network based approach is commonly employed for this process and Convolutional Neural Network which is a Deep Neural Network method, is very popular. Haar cascade classifier is used for face detection. A publicly available database is used to test the architecture of the proposed Convolutional Neural Network.