Research Article
[Retracted] Multidimensional CNN-Based Deep Segmentation Method for Tumor Identification
Table 1
2D-ResUNet and 3D-ResUNet network structures.
| Network layer | 2D-ResUNet | 3D-ResUNet | Feature map size | Network layer size | Feature map size | Network layer size |
| Input | | — | | — | Residual structure 1 | | | | | Max pooling layer 1 | | max pooling | | max pooling | Residual structure 2 | | | | | Max pooling layer 2 | | max pooling | | max pooling | Residual structure 3 | | | | | Max pooling layer 3 | | max pooling | | max pooling | Residual structure 4 | | | | | Max pooling layer 4 | | max pooling | | max pooling | Residual structure 5 | | | | | Deconvolution 1 | | , -[residual structure 4] | | , -[residual structure 4] | Deconvolution 2 | | , -[residual structure 3] | | , -[residual structure 3] | Deconvolution 3 | | , -[residual structure 2] | | , -[residual structure 2] | Deconvolution 4 | | , -[residual structure 1] | | , -[residual structure 1] | Convolutional layer | | , 2 | | , 2 |
|
|