Lesion segmentation on clinical breast ultrasound dataset
92.17mDSCRad-VLSM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Rad-VLSMParadigm=Weakly-supervised/Foundation models2026.05 | 92.17 | 85.79 | |
| MedCLIP-SAMv2Paradigm=Weakly-supervised/Foundation models2026.05 | 88.65 | 81.69 | |
| BoxShrinkParadigm=Weakly-supervised/Foundation models2026.05 | 87.35 | 80.39 | |
| WeakPolypParadigm=Weakly-supervised/Foundation models2026.05 | 85.12 | 75.15 | |
| SimTxtSegParadigm=Weakly-supervised/Foundation models2026.05 | 84.03 | 74.44 | |
| ARSegParadigm=Weakly-supervised/Foundation models2026.05 | 80.88 | 68.75 | |
| EMCADParadigm=Transformer-based2026.05 | 74.62 | 69.19 | |
| PVT-CascadeParadigm=Transformer-based2026.05 | 73.24 | 67.15 | |
| TransUNetParadigm=Transformer-based2026.05 | 71.98 | 65.51 | |
| Swin-UNetParadigm=Transformer-based2026.05 | 66.29 | 60.35 | |
| Att-UNetParadigm=Mainstream encoder-decoder2026.05 | 65.58 | 60.54 | |
| UNetParadigm=Mainstream encoder-decoder2026.05 | 65.45 | 60.35 | |
| S2MEParadigm=Weakly-supervised/Foundation models2026.05 | 61.15 | 45.14 |