Skip to main content

Table 10 Architecture of RN

From: Reverse active learning based atrous DenseNet for pathological image classification

Layer Type Kernel size & number
1 C 3×3,16
2 MP 2×2
3 C 3×3,32
4 MP 2×2
5 C 3×3,64
6 MP 2×2
7 C 3×3,64
8 MP 2×2
9 C 3×3,128
10 MP 2×2
11 C 3×3,128
12 AP 7×7
13 FC 256
14 FC 4
  1. Pipeline consists of convolution layer(C), max pooling layer(MP), average pooling layer(AP) and fully-connected layer(FC)