Research on Facial Expression Recognition Based on Deep Learning Method

Facial expression recognition has been a heated investigate topic during recent years which can be applied in various aspects of people’s daily life, such as entertainment pictures’ generation, check-in machine of a company, etc. In this work we investigate the effect of adding visual attention mechanism to a facial expression recognition deep learning method. The main contribution is to form a new architecture to apply SE Block to normal VGGNet which shows that a significant improvement on the accuracy of facial expression recognition. These findings are properly tested on 66.8% accuracy, which validates the effectiveness of the proposed SE-VGGNet.