Video Based Deep CNN Model for Depression Detection

Abstract

Our face reflects our feelings towards anything and everything we see, smell, teste or feel through any of our senses. Hence multiple attempts have been made since last few decades towards understanding the facial expressions. Emotion detection has numerous applications since Safe Driving, Health Monitoring Systems, Marketing and Advertising etc. We propose an Automatic Depression Detection (ADD) system based on Facial Expression Recognition (FER).
 We propose a model to optimize the FER system for understanding seven basic emotions (joy, sadness, fear, anger, surprise, disgust and neutral) and use it for detection of Depression Level in the subject. The proposed model will detect if a person is in depression and if so, up to what extent. Our model will be based on a Deep Convolution Neural Network (DCNN).