Table 2.
Quantitative comparison of the proposed model and the state-of-the-art. CBAM named the model to which CBAM was applied just before each pooling in U-Net. Except for the proposed method, other networks applied instance normalization
| DSC | HD | Precision | Recall | |
|---|---|---|---|---|
| CE-Net | 0.664 0.314 |
20.483 21.176 |
0.691 0.315 |
0.728 0.346 |
| CBAM | 0.671 0.298 |
20.961 22.387 |
0.709 0.274 |
0.755 0.329 |
| U-Net | 0.672 0.308 |
17.832 21.422 |
0.722 0.323 |
0.702 0.332 |
| CA-Net | 0.672 0.336 |
13.288
17.229
|
0.737 0.216 |
0.683 0.353 |
| MSAM | 0.703 0.291 |
14.384 18.470 |
0.742 0.287 |
0.742 0.302 |
| Attention U-Net | 0.718 0.250 |
15.538 17.604 |
0.733 0.275 |
0.764 0.252 |
| Proposed |
0.731
0.228
|
13.393 15.813 |
0.756
0.257
|
0.768
0.236
|



























