Skip to main content
. 2019 Oct 17;13:67. doi: 10.3389/fninf.2019.00067

Table 2.

The MeshNet dilated convolutional neural network architecture used for brain segmentation.

Layer Filter Padding Dilation (l) Non-linearity
1 96 × 33 1 1 ReLU
2 96 × 33 1 1 ReLU
3 96 × 33 1 1 ReLU
4 96 × 33 2 2 ReLU
5 96 × 33 4 4 ReLU
6 96 × 33 8 8 ReLU
7 96 × 33 1 1 ReLU
8 50 × 13 0 1 Softmax