Skip to main content
. 2019 Sep 30;6:193. doi: 10.3389/fmed.2019.00193

Figure 9.

Figure 9

CNN architecture overview. A convolution layer with 16 3 × 3 filters, and a stride of 1 × 1 is denoted by n16s1. The same, followed by a batch normalization layer (yellow) and a ReLU non-linearity (green) is denoted by n16_bn_relu. The learning rate used for training is 0.0001, the optimizer is SGD with momentum 0.9 and the batch size is set to 16.