Fig. 5
From: GEM-CRAP: a fusion architecture for focal seizure detection
The three-dimensional feature topography in the convolutional layers is composed of frequency steps, channel counts, and activation values. It encompasses a comparison of the three-layer convolution process and the CNN-RES feature extraction performance. (1) [conv1-conv2-conv3] (2) [conv1-conv2-RES2] (3) [conv1-RES1-RES2]