Research On Facial Expression Recognition Technology Based On Deep Convolutional Neural Network

Facial expression recognition has broad application prospects in fields such as safe driving,medical care,education,and commercial sales.The study of fast and accurate recognition methods for facial expressions will help provide necessary technical support for facial expression recognition in different fields.The facial expression recognition algorithm based on deep convolutional neural network has higher recognition accuracy than traditional facial expression recognition algorithms.However,the deep convolutional neural network model has many parameters and complex structure,which leads to slow model training and the process of self-learning to extract facial expression features.There is a certain degree of arbitrariness in the model,which leads to problems such as insufficient expressiveness of facial expression features extracted by the model.These problems seriously affect the recognition effect of the algorithm model.In response to these issues,this article has conducted an in-depth study,this article has conducted an in-depth study.First of all,this paper designs an expression recognition algorithm based on grouped recombination residual network.This algorithm uses grouped convolution instead of standard convolution to extract facial expression feature information.In order to effectively fuse the expression feature information output by different groups,channel reorganization is added to the grouped convolution,and different levels of expressions are transmitted through the cross-connection of the residual network.Features realize the transfer and fusion of expression features between different levels of the convolutional network.The algorithm in this paper effectively reduces the number of parameters of the network model,prevents the overall performance degradation of the network model during training,and improves the training speed of the network And recognition accuracy.Secondly,this paper designs a grouped residual network expression recognition algorithm based on multi-attention mechanism.Aiming at the problem that the expression features extracted by the convolutional neural network are not very expressive,the channel and spatial attention mechanism fusion algorithm is used to effectively extract the expression features with strong expressiveness in the channel and spatial dimensions,while suppressing the weak expressive expressions feature.Thereby improving the accuracy of facial expression recognition.The research in this paper shows that the use of grouped recombination residual network can effectively reduce the number of network model parameters,increase the training speed of convolutional network expression recognition,and improve the accuracy of expression recognition of the network model.The multi-attention fusion mechanism can extract the eigenfeatures of facial expressions more effectively.The experimental results show the effectiveness of the two expression recognition algorithms designed in this paper in improving the training speed of the model and the accuracy of expression recognition.