An Improved Method for Facial Expression Recognition Based on Attention Mechanism
Targeting the problems of poor recognition accuracy and a large number of Deep Learning model parameters due to light and posture influence in facial expression recognition,this paper proposes an improved Convolutional Neural Network model based on Attention Mechanism.Through the introduction of the Attention Mechanism module,the model selectively focuses on the locally important information of the target object and reduces the interference of irrelevant information,while using a neural network with fewer neurons and a large convolutional kernel,the parameters of the network are significantly decreased,and the method builds a lightweight Convolutional Neural Network model with a shallower hierarchy and fewer parameters.Experiments are conducted on the CK+facial expression dataset,and results show that the proposed method significantly reduces model parameters while ensuring facial recognition accuracy,with an accuracy rate of 96.37%.
国家科技期刊平台
NSTL
万方数据
维普
