Expression-EEG Bimodal Fusion Emotion Recognition Method Based on Deep Learning

Table 1

Experimental simulation results of fine tuning of each model.

Network model	Pretraining data set	FER 2013 test set results (%)	SFEW test set results
Baseline DCNN	Null	56.21	54.35
GoogLeNet	ImageNet	61.23	59.21
CaffeNet	ImageNet	67.41	56.22
Residual network	ImageNet	72.34	68.34
VGG	ImageNet	76.21	71.23
VGG-FACE	Faces	89.21	78.24