Skin Lesion Segmentation on ISIC 2016 (test)
94.31mDiceSiamese-Diffusion
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Siamese-DiffusionArchitecture=SegFormer, Training Set=Real + Synthetic, Mask Prior=Random2025.05 | 94.31 | 90.01 | |
| Siamese-DiffusionArchitecture=SegFormer, Training Set=Real + Synthetic, Mask Prior=Real2025.05 | 94.14 | 89.76 | |
| T2I-AdapterArchitecture=SegFormer, Training Set=Real + Synthetic2025.05 | 93.71 | 89.24 | |
| Real DatasetArchitecture=SegFormer, Training Set=Real Only2025.05 | 93.57 | 89.06 | |
| ControlNetArchitecture=SegFormer, Training Set=Real + Synthetic2025.05 | 93.54 | 88.83 | |
| Copy-PasteArchitecture=SegFormer, Training Set=Real + Synthetic2025.05 | 93.53 | 88.85 | |
| Siamese-DiffusionArchitecture=UNet, Training Set=Real + Synthetic, Mask Prior=Random2025.05 | 90.06 | 87.25 | |
| Siamese-DiffusionArchitecture=UNet, Training Set=Real + Synthetic, Mask Prior=Real2025.05 | 89.91 | 87.01 | |
| Real DatasetArchitecture=UNet, Training Set=Real Only2025.05 | 89.58 | 86.62 | |
| T2I-AdapterArchitecture=UNet, Training Set=Real + Synthetic2025.05 | 89.48 | 86.47 | |
| Copy-PasteArchitecture=UNet, Training Set=Real + Synthetic2025.05 | 89.43 | 86.48 | |
| ControlNetArchitecture=UNet, Training Set=Real + Synthetic2025.05 | 89.42 | 86.51 |