Skip to main content
. 2020 Jun;6(2):250–260. doi: 10.18383/j.tom.2020.00003

Figure 1.

Figure 1.

Visual geometry group-slow (VGG-S) architecture containing 5 convolution layers, conv 1–conv 5, and 3 fully connected layers. A × K × K means subimages created using a K × K kernel; st. indicates convolution stride; LRN indicates local response normalization; pad indicates padding; pool indicates max pooling; and ReLU indicates rectified linear units.