Deep Learning Based Facial Expression Recognition And Face Recognition Model

Face Expression is a very important aspect of human intelligence.Among other words,the most convincing means of expressing human emotion is through facial expressions.There has recently been a trend towards deep-learning,and especially Convolution Neural Network has been actively researched upon in the last decade due to its high number of applications in the human-computer interaction domain.Researchers have been able to obtain good results on images taken in laboratories under the same conditions,or at least similar.However,on more arbitrary face pictures with various head poses and lighting these models don't work very well.Wild expressions are captured in a real-world setting and have natural expressions.This thesis aims to explain the challenges facial expression recognition(FER)is facing in this wild environment.This thesis contributions are in three parts.First,I analyze three popular architectures of the neural network,and their effectiveness in classifying facial expressions.Secondly,I am presenting two modifications to one of these networks which lead to the proposed STN-COV model.Third,Explain my research paper recently published paper title: “Unsupervised Facial Expression Detection Using Genetic Algorithm”.