Table 3.
Performance comparison between the proposed system and other works.
| Method/Dimension | Reference/Year | Liver | Lesion |
|---|---|---|---|
| residual UNet/2.5D | [18]/2017 | - | 67% |
| UNet + graph cut/2D | [13]/2019 | 95.05% | - |
| TDP-CNN/3D | [14]/2020 | 96.50% | 68.9% |
| 3D RA-UNet/3D | [16]/2020 | 96.10% | - |
| SAR-UNet/2D | [15]/2021 | 95.71% | 59.5% |
| 2.5D UNet/similar3D | [19]/2021 | 95.30% | 78.4 ± 16% |
| XNet/3D | [20]/2021 | 97.10% | 84.3% |
| 3D U-Net/3D | [34]/2022 | 94.20% | |
| CEDRNN/2D | [35]/2022 | 95.20% | - |
| EfficientNet and Attention-Based Residual U-Net/2D |
[36]/2022 | 95.20% | - |
| Decision fusion/2D | our | 95.76% | 75.83% |