Research Article
A Collaborative Dictionary Learning Model for Nasopharyngeal Carcinoma Segmentation on Multimodalities MR Sequences
Table 2
Architecture of the FCN network for tumor localization.
| | Type | Input size | Output size | Filter size | Stride | # filters |
| Layer 1 | Conv. | | | | | 32 | Layer 2 | Max-pool. | | | | | ā | Layer 3 | Conv. | | | | | 64 | Layer 4 | Max-pool. | | | | | ā | Layer 5 | Conv. | | | | | 128 | Layer 6 | Max-pool. | | | | | ā | Layer 7 | Conv. | | | | | 128 | Layer 8 | Conv. | | | | | 128 | Layer 9 | Conv. | | | | | 128 | Layer 10 | Conv. | | | | | 128 | Layer 11 | Upsampling | | | | | ā | Layer 12 | Conv. | | | | | 128 | Layer 13 | Upsampling | | | | | ā | Layer 14 | Conv. | | | | | 64 | Layer 15 | Upsampling | | | | | ā | Layer 16 | Conv. | | | | | 32 | Layer 17 | Conv. | | | | | 2 |
|
|
The convolutional layer is denoted by Conv., and the max pooling by max-pool. |