Table 2.
Network Layer Architecture of CTANet.
| Layer Name | Input Size | Output Size | Filters (Channels) |
|---|---|---|---|
| Conv Stem | (512, 512, 3) | (128, 128, 96) | 96 |
| Downsample-1 | (128, 128, 96) | (64, 64, 192) | 192 |
| Downsample-2 | (64, 64, 192) | (32, 32, 384) | 384 |
| Downsample-3 | (32, 32, 384) | (16, 16, 768) | 768 |
| ConvNeXt-1 | (128, 128, 96) | (128, 128, 96) | 96 |
| ConvNeXt-2 | (64, 64, 192) | (64, 64, 192) | 192 |
| ConvNeXt-3 | (32, 32, 384) | (32, 32, 384) | 384 |
| ConvNeXt-4 | (16, 16, 768) | (16, 16, 768) | 768 |
| MSM-AMM-1 | (128, 128, 96) | (128, 128, 96) | 96 |
| MSM-AMM-2 | (64, 64, 192) | (64, 64, 192) | 192 |
| MSM-AMM-3 | (32, 32, 384) | (32, 32, 384) | 384 |
| BHAFormer-1 | (128, 128, 96) | (128, 128, 96) | 96 |
| BHAFormer-2 | (64, 64, 192) | (64, 64, 192) | 192 |
| BHAFormer-3 | (32, 32, 384) | (32, 32, 384) | 384 |
| Upsample-1 | (16, 16, 768) | (32, 32, 384) | 384 |
| Upsample-2 | (32, 32, 384) | (64, 64, 192) | 192 |
| Upsample-3 | (64, 64, 192) | (128, 128, 96) | 96 |
| Upsample-4 | (128, 128, 96) | (512, 512, 96) | 96 |
| Segment Head | (512, 512, 96) | (512, 512, 2) | 2 (Classes) |