A VISUAL ATTENTION BASED VGG19 NETWORK FOR FACIAL EXPRESSION RECOGNITION

Abstract

Facial emotion recognition (FER) is meaningful for human-machine interaction such as clinical practice, playing games, and behavioral description. FER has been an active area of research over the past few decades, and it is still challenging due to the high intra-class variation, the heterogeneity of human faces, and variations in images such as different facial poses and various lighting conditions. Recently, deep learning models have shown great potential for FER. Besides, the visual attention technique has helped deep learning networks improve. In this paper, we present a visual attention-based VGG19 network for FER. The proposed outperforms the state-of-the-art methods slightly on the FER2013 dataset.