Aiming at the problems of insufficient extraction of facial expression features by a single convolutional neural network and large amount of parameters

a facial expression recognition algorithm fused with parallel network features is proposed. The algorithm first modifies the residual block in the ResNet network

reduces the amount of network parameters and uses pre-activation to reduce the error rate. After that

the features extracted by the improved ResNet network and the features extracted by the VGG network after the cut layer are merged to obtain the network model P-ResNet-VGG

in which the loss function uses the cross-entropy loss function. This model has been extensively tested on the FER2013 and JAFFE datasets. Experimental results show that this model has improved accuracy on FER2013 and JAFFE expression data sets than other models

and has better robustness.