Research Article
[Retracted] Rethinking Separable Convolutional Encoders for End-to-End Semantic Image Segmentation
Table 6
Structure details of the generated network.
| Operation | Convolution kernel size | Input channel | Output channel | Step | Batch normalization |
| Conv 1 | 12∗12 | 12 | 144 | 1 | N | Conv 2 | 4∗4 | 144 | 144 | 1 | Y | Conv 3 | 4∗4 | 144 | 144 | 1 | Y | Conv 4 | 4∗4 | 144 + 144 | 144 | 1 | Y | Conv 5 | 4∗4 | 144 + 144 | 432 | 1 | N | Conv 6 | 12∗12 | 432 | 4 | 1 | N |
|
|