Facial Expression Recognition Based on VGGNet Convolutional Neural Network

Abstract

Aiming at the low recognition rate of traditional convolutional neural network in facial expression database, I propose a facial expression recognition method based on VGGNet deep convolutional neural network. With a deeper network architecture and a 3*3 small convolution kernel and a 2*2 small pool kernel, the recognition rate is significantly improved, and the number of parameters is only slightly larger than that of the shallow layer. In order to further reduce the number of parameters, only the first fully-connected layer of the original network is retained; in order to prevent over-fitting, the data set is multiple croped and dropout strategy is introduced before the fully-connected layer. Finally, the Softmax classifier is used for classification and recognition in the network. Experimental results show that the recognition rate of the algorithm in FER2013 database is 73.06%.