Automatic Semantic Segmentation of Brain Gliomas from MRI Images Using a Deep Cascaded Neural Network

Table 4

A list of parameters used in the proposed subnet ITCN. In each convolutional layer, the feature maps had been padded by 1 prior to the convolution so that the convolution do not change the size of the resultant feature map.

Number	Layer name	Filter size	Stride	Number of filters	FC units	Output
1	Conv 1_1 + LReLU	33	1	64	—	333364
2	Conv 1_2 + LReLU	33	1	64	—	333364
3	Conv 1_3 + LReLU	33	1	64	—	333364
4	Max pooling 1	33	2	—	—	161664
5	Conv 2_1 + LReLU	33	1	128	—	1616128
6	Conv 2_2 + LReLU	33	1	128	—	1616128
7	Conv 2_3 + LReLU	33	1	128	—	1616128
8	Max pooling 2	33	2	—	—	88128
9	FC1 + dropout	—	—	—	8192	256
10	FC2 + dropout	—	—	—	256	128
11	FC3 + softmax	—	—	—	128	4