Table 3. To evaluate the generalization ability of different segmentation techniques, use the Local data set to test after training on the York dataset, and the results are presented in the form of average (std.).
| Local | ||||
|---|---|---|---|---|
| Dice index | HD (mm) | |||
| LV | MYO | LV | MYO | |
| U-Net | 0.69 (0.09) | 0.85 (0.07) | 22.51 (6.58) | 21.55 (8.41) |
| FC-DenseNet | 0.74 (0.07) | 0.89 (0.06) | 18.93 (6.87) | 16.74 (5.88) |
| NCDN | 0.78 (0.09) | 0.89 (0.06) | 14.94 (6.32) | 17.13 (6.84) |
Notes.
Optimal values are indicated in bold.