Research on Facial Microexpression Classification Based on Graph Attention Neural Network
In this paper,a method of facial microexpression classification based on graph attention neural network is proposed.In this method,the image convolutional neural network is mainly used to extract features from the microexpression images,and the attention mechanism is introduced to combine with the image neural network to assign different weights to different nodes in the image to achieve more effective microexpression classification.The results show that the recognition rate of each microexpres-sion is improved by using the graph attention neural network method.Among them,the recognition rate of happy expression is as high as 87.52％,which is 7.52％higher than that of GCN.Meanwhile,the fear recognition rate of this method is 63.27％,which is 8.61％higher than that of GCN.It shows that the graph attention neural network method which introduces attention mechanism has higher preci-sion of micro-expression recognition.
万方数据
维普
