Optimizing Residual Networks and VGG for Classification of EEG Signals: Identifying Ideal Channels for Emotion Recognition

Table 3

Model training convergence efficiency comparison between ResNet and VGG.

	Training length to reach 95% training accuracy (epochs//seconds)
	Using all 62 channels	Using outermost 10 channels
ResNet18-1D (T-then-s)	11//1665	11//416
VGG14-1D (T-then-s)	10//2488	10//503
VGG16-1D (T-then-s)	14//3505	12//622