Facial Expression Recognition With Convolutional Neural Networks

Abstract

In this project,development of convolutional neural networks (CNN) for a facial expression recognition has been performed.The main objective involves segregation of each facial image into one of the seven facial emotion categories contemplated in this study from the fer2013 dataset.The process consists of preparing the dataset into a form which will work on a generalized algorithm,followed by face detection stage,in which the face is detected from the image that we wish to classify and then implementing the CNN algorithm to segregate input image into one of seven classes.We experimented with different models such as VGG16,VGG19 and ResNet50,achieving different accuracy for different models.