Depression Detection from Facial Behaviour through Deep Learning

Abstract

Millions of people worldwide suffer from depression. There are some differences in condition of mental health between two people who have the same disorder. The degree of depression is analyzed through video-recorded clinical meetings. Depression is a mood disorder and is a challenging issue in now days. In worldwide there are 350 million people suffering from depression. Depression patients find hard to concentrate on their work. They may suffer from insomnia, restlessness, loss of appetite and sometimesthey also get suicidal thoughts. In this paper we discuss about depression detection methods. This paper elaborates how Hand-Crafted Method, Deep Convolutional Neural Network, Raw Audio, Spectrogram Audio, Local Binary Pattern, Median Robust Extended Local Binary Pattern, Joint Tuning, Naïve Bayes Algorithm, Gaussian Mixture Model etc. are used in depression detection. Here we discuss some audio, video and text recognition methods and try to find best detection technique for depression.