Table 2 Layer structure of Network 2
From: Artifact estimation network for MR images: effectiveness of batch normalization and dropout layers
Layer no. | Structure | Layer no. | Structure | Layer no. | Structure |
|---|---|---|---|---|---|
1 | Batch normalization 1 | 13 | Dropout 3 (0.2) | 24–1 | Conv 7 |
2–1 | Conv 1 | 14–1 | Conv 4 | 24–2 | Conv 7 |
2–2 | Conv 1 | 14–2 | Conv 4 | 25 | Upsampling |
3 | Max pooling 1 | 15 | Max pooling 4 | 26 | Batch normalization 8 |
4 | Batch normalization 2 | 16 | Batch normalization 5 | 27 | Dropout 6 (0.1) |
5 | Dropout 1 (0.1) | 17 | Dropout 4 (0.2) | 28–1 | Conv 8 |
6–1 | Conv 2 | 18–1 | Conv 5 | 28–2 | Conv 8 |
6–2 | Conv 2 | 18–2 | Conv 5 | 29 | Upsampling |
7 | Max pooling 2 | 19 | Upsampling | 30 | Batch normalization 9 |
8 | Batch normalization 3 | 20 | Batch normalization 6 | 31 | Dropout 7 (0.1) |
9 | Dropout 2 (0.1) | 21 | Dropout 5 (0.2) | 32–1 | Conv 9 |
10–1 | Conv 3 | 21–1 | Conv 6 | 32–2 | Conv 9 |
10–2 | Conv 3 | 213 | Conv 6 | 33 | Conv 10 |
11 | Maxpooling 3 | 22 | Upsampling | 34 | Regression (output) |
13 | Batch normalization 4 | 23 | Batch normalization 7 | Â | Â |