Skin Lesion Segmentation on ISIC 2017 (test)
91.4Dice ScoreOurs
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Ours2025.12 | 91.4 | — | 92.75 | 97.78 | 97.26 | — | — | — | — | |
| LKA2025.12 | 90.99 | — | 90.55 | 98.49 | 96.98 | — | — | — | — | |
| Hiformer-B2025.12 | 90.93 | — | 88.67 | 98.57 | 96.69 | — | — | — | — | |
| UltraLightVM-UNet2025.12 | 90.91 | — | 90.53 | 97.9 | 96.46 | — | — | — | — | |
| TBConvL-Net2025.12 | 90.89 | — | 91.19 | 97.61 | 96.07 | — | — | — | — | |
| Samba+2026.02 | 90.65 | — | — | 98.05 | 96.86 | 82.9 | — | — | — | |
| VM-UNetV2year=20242026.02 | 90.31 | — | — | 97.67 | 96.7 | 82.34 | — | — | — | |
| UNetV2year=20252026.02 | 90.22 | — | — | 97.94 | 96.78 | 82.18 | — | — | — | |
| EMCAD2025.12 | 90.06 | — | 93.7 | 96.81 | 96.55 | — | — | — | — | |
| UGDD-Net (Ours)Backbone=Dual-Domain, Architecture Type=Dual-Path, Mechanism=Spatial-Frequency (SF) + Uncertainty-Aware (U)2026.05 | 89.35 | — | — | — | — | 81.97 | 12.68 | 4.72 | 10.92 | |
| EGE-Unet2025.12 | 89.28 | — | 88.39 | 98.11 | 96.48 | — | — | — | — | |
| VM-UNetyear=20242026.02 | 89.03 | — | — | 97.58 | 96.29 | 80.23 | — | — | — | |
| GCA-ResUNet2025.12 | 88.93 | — | 89.9 | 98.58 | 96.81 | 79.91 | — | — | — | |
| UltraLBM-UNetVenue (year)=-, Params (M)=0.034, FLOPs (G)=0.062025.12 | 88.78 | — | — | — | — | 79.82 | — | — | — | |
| EGE-UNetyear=20232026.02 | 88.77 | — | — | 97.62 | 96.19 | 79.81 | — | — | — | |
| TransFuseyear=20212026.02 | 88.4 | — | — | 97.98 | 96.17 | 79.21 | — | — | — | |
| TransFuse2025.12 | 88.4 | — | 87.14 | 97.98 | 96.17 | 79.21 | — | — | — | |
| Swin-Unet2025.12 | 88.15 | — | 83.64 | 98.69 | 95.82 | — | — | — | — | |
| MALUNet2025.12 | 88.13 | — | 84.78 | 98.47 | 96.18 | 78.78 | — | — | — | |
| MALUNetVenue (year)=BIBM (2022), Params (M)=0.178, FLOPs (G)=0.0832025.12 | 88.09 | — | — | — | — | 78.71 | — | — | — | |
| MK-UNetVenue (year)=ICCVW (2025), Params (M)=0.316, FLOPs (G)=0.3282025.12 | 88 | — | — | — | — | 78.57 | — | — | — | |
| UltraLBM-UNet-TVenue (year)=-, Params (M)=0.011, FLOPs (G)=0.0192025.12 | 88 | — | — | — | — | 78.57 | — | — | — | |
| EGE-UNetVenue (year)=MICCAI (2023), Params (M)=0.053, FLOPs (G)=0.0722025.12 | 87.84 | — | — | — | — | 78.32 | — | — | — | |
| AD-LA Former2025.12 | 87.68 | — | 93.18 | 98.65 | 97.03 | — | — | — | — | |
| U-LiteVenue (year)=APSIPA (2023), Params (M)=0.878, FLOPs (G)=0.7572025.12 | 87.61 | — | — | — | — | 77.95 | — | — | — | |
| UltraLight VM-UNetVenue (year)=Patterns (2024), Params (M)=0.045, FLOPs (G)=0.0692025.12 | 87.59 | — | — | — | — | 77.93 | — | — | — | |
| TinyU-NetVenue (year)=MICCAI (2024), Params (M)=0.481, FLOPs (G)=1.6592025.12 | 87.57 | — | — | — | — | 77.89 | — | — | — | |
| CMUNeXt-SVenue (year)=ISBI (2024), Params (M)=0.418, FLOPs (G)=1.092025.12 | 87.56 | — | — | — | — | 77.87 | — | — | — | |
| UNeXtVenue (year)=MICCAI (2022), Params (M)=1.472, FLOPs (G)=0.5732025.12 | 87.52 | — | — | — | — | 77.81 | — | — | — | |
| U-KANVenue (year)=AAAI (2025), Params (M)=25.359, FLOPs (G)=6.8892025.12 | 87.46 | — | — | — | — | 77.71 | — | — | — | |
| Rolling-UNet-SVenue (year)=AAAI (2024), Params (M)=1.783, FLOPs (G)=2.1022025.12 | 87.38 | — | — | — | — | 77.59 | — | — | — | |
| UDELBackbone=CNN+Mamba, Architecture Type=Dual-Path, Mechanism=Uncertainty-Aware (U)2026.05 | 87.34 | — | — | — | — | 79.57 | 14.61 | 5.96 | 13.47 | |
| UTNetV22025.12 | 87.23 | — | 84.85 | 98.05 | 95.84 | 77.35 | — | — | — | |
| MambaU-LiteVenue (year)=ICISN (2025), Params (M)=0.416, FLOPs (G)=0.9342025.12 | 87.21 | — | — | — | — | 77.32 | — | — | — | |
| Attention U-NetVenue (year)=MIDL (2018), Params (M)=34.879, FLOPs (G)=66.6322025.12 | 87.02 | — | — | — | — | 77.02 | — | — | — | |
| U-Net2025.12 | 86.99 | — | 86.82 | 97.43 | 95.65 | 76.98 | — | — | — | |
| SFma-UNetBackbone=Mamba, Architecture Type=Dual-Path, Mechanism=Spatial-Frequency (SF)2026.05 | 86.91 | — | — | — | — | 78.99 | 15.38 | 6.25 | 16.92 | |
| UNet++Venue (year)=DLMIA (2018), Params (M)=9.163, FLOPs (G)=34.9032025.12 | 86.77 | — | — | — | — | 76.64 | — | — | — | |
| CDMT-UNetBackbone=CNN+Trans, Architecture Type=Single-Path, Mechanism=Spatial-Frequency (SF)2026.05 | 86.26 | — | — | — | — | 78.22 | 15.65 | 6.32 | 18.65 | |
| U-NetVenue (year)=MICCAI (2015), Params (M)=7.766, FLOPs (G)=13.7462025.12 | 86.25 | — | — | — | — | 75.84 | — | — | — | |
| MPBA-NetBackbone=CNN+Trans, Architecture Type=Single-Path2026.05 | 86.17 | — | — | — | — | 78.11 | 16.93 | 6.82 | 17.57 | |
| F2CAU-NetBackbone=CNN, Architecture Type=Dual-Path, Mechanism=Uncertainty-Aware (U)2026.05 | 86.16 | — | — | — | — | 78.15 | 16.28 | 6.62 | 14.11 | |
| WinGraphUNetBackbone=CNN+Graph, Architecture Type=Single-Path2026.05 | 86.1 | — | — | — | — | 77.94 | 14.74 | 6.02 | 17.84 | |
| DPMF-NetBackbone=CNN+Trans, Architecture Type=Dual-Path, Mechanism=Spatial-Frequency (SF)2026.05 | 85.96 | — | — | — | — | 77.85 | 15.84 | 6.45 | 16.14 | |
| DPGNetBackbone=CNN+Trans, Architecture Type=Single-Path, Mechanism=Uncertainty-Aware (U)2026.05 | 85.68 | — | — | — | — | 77.34 | 16.98 | 6.86 | 14.38 | |
| DEviSBackbone=CNN, Architecture Type=Single-Path, Mechanism=Uncertainty-Aware (U)2026.05 | 85.43 | — | — | — | — | 77.16 | 17.45 | 7.15 | 13.91 | |
| AHF-U-NetBackbone=CNN, Architecture Type=Single-Path, Mechanism=Uncertainty-Aware (U)2026.05 | 85.38 | — | — | — | — | 77.02 | 17.29 | 6.91 | 14.76 | |
| CFFormerBackbone=CNN+Trans, Architecture Type=Dual-Path2026.05 | 85.32 | — | — | — | — | 76.97 | 17.69 | 7.2 | 18.15 | |
| SkinMambaBackbone=CNN+Mamba, Architecture Type=Single-Path, Mechanism=Spatial-Frequency (SF)2026.05 | 85.26 | — | — | — | — | 76.88 | 15.63 | 6.37 | 16.27 | |
| DSU-NetBackbone=CNN+Trans, Architecture Type=Dual-Path2026.05 | 85.25 | — | — | — | — | 76.9 | 18.59 | 7.24 | 18.04 | |
| DMA-NetBackbone=CNN+Mamba, Architecture Type=Dual-Path2026.05 | 85.06 | — | — | — | — | 76.74 | 17.57 | 7.16 | 16.38 | |
| SF-UNetBackbone=CNN, Architecture Type=Dual-Path, Mechanism=Spatial-Frequency (SF)2026.05 | 84.95 | — | — | — | — | 76.64 | 16.07 | 6.48 | 15.94 | |
| VM-UNetBackbone=Mamba, Architecture Type=Single-Path2026.05 | 84.87 | — | — | — | — | 76.38 | 18.92 | 7.31 | 16.26 | |
| BDFormerBackbone=Trans, Architecture Type=Dual-Path2026.05 | 84.85 | — | — | — | — | 76.42 | 16.38 | 6.69 | 19.44 | |
| H-NetBackbone=CNN, Architecture Type=Dual-Path2026.05 | 84.54 | — | — | — | — | 75.98 | 19.03 | 7.35 | 15.87 | |
| U-NetBackbone=CNN, Architecture Type=Single-Path2026.05 | 83.87 | — | — | — | — | 74.94 | 19.11 | 7.42 | 15.77 | |
| Swin-UnetBackbone=Trans, Architecture Type=Single-Path2026.05 | 83.46 | — | — | — | — | 74.63 | 19.85 | 7.68 | 19.79 | |
| NCE+RCEalpha=0.02023.06 | 82.9 | — | — | — | — | — | — | — | — | |
| GCEalpha=0.02023.06 | 82.8 | — | — | — | — | — | — | — | — | |
| SCEalpha=0.02023.06 | 82.8 | — | — | — | — | — | — | — | — | |
| NGCE+MAEalpha=0.02023.06 | 82.8 | — | — | — | — | — | — | — | — | |
| RCEalpha=0.02023.06 | 82.7 | — | — | — | — | — | — | — | — | |
| NGCE+RCEalpha=0.02023.06 | 82.7 | — | — | — | — | — | — | — | — | |
| MAEalpha=0.02023.06 | 82.6 | — | — | — | — | — | — | — | — | |
| NGCEalpha=0.02023.06 | 82.5 | — | — | — | — | — | — | — | — | |
| T-Lossalpha=0.02023.06 | 82.5 | — | — | — | — | — | — | — | — | |
| TransUNet2025.12 | 81.23 | — | 82.63 | 95.77 | 92.07 | — | — | — | — | |
| BPC-NetLearning Paradigm=Unsupervised2026.04 | 81.03 | — | 86.45 | 94.08 | 91.83 | 72.05 | — | — | — | |
| T-Lossalpha=0.3, beta=0.52023.06 | 80.9 | — | — | — | — | — | — | — | — | |
| NGCE+RCEalpha=0.3, beta=0.52023.06 | 80.7 | — | — | — | — | — | — | — | — | |
| SCEalpha=0.3, beta=0.52023.06 | 80.6 | — | — | — | — | — | — | — | — | |
| GCEalpha=0.3, beta=0.52023.06 | 80.5 | — | — | — | — | — | — | — | — | |
| USL-NetLearning Paradigm=Unsupervised2026.04 | 80.45 | — | 88.59 | 93.68 | 90.47 | 68.45 | — | — | — | |
| T-Lossalpha=0.3, beta=0.72023.06 | 80.4 | — | — | — | — | — | — | — | — | |
| MAEalpha=0.3, beta=0.52023.06 | 80.3 | — | — | — | — | — | — | — | — | |
| NGCEalpha=0.3, beta=0.52023.06 | 80.3 | — | — | — | — | — | — | — | — | |
| RCEalpha=0.3, beta=0.52023.06 | 80.2 | — | — | — | — | — | — | — | — | |
| NGCE+MAEalpha=0.3, beta=0.52023.06 | 80.2 | — | — | — | — | — | — | — | — | |
| T-Lossalpha=0.5, beta=0.52023.06 | 80 | — | — | — | — | — | — | — | — | |
| NCE+RCEalpha=0.3, beta=0.52023.06 | 79.9 | — | — | — | — | — | — | — | — | |
| SCEalpha=0.3, beta=0.72023.06 | 79.3 | — | — | — | — | — | — | — | — | |
| NCE+RCEalpha=0.3, beta=0.72023.06 | 79.2 | — | — | — | — | — | — | — | — | |
| RCEalpha=0.3, beta=0.72023.06 | 79.1 | — | — | — | — | — | — | — | — | |
| NGCE+RCEalpha=0.3, beta=0.72023.06 | 79 | — | — | — | — | — | — | — | — | |
| T-Lossalpha=0.5, beta=0.72023.06 | 79 | — | — | — | — | — | — | — | — | |
| NGCEalpha=0.3, beta=0.72023.06 | 78.8 | — | — | — | — | — | — | — | — | |
| NGCE+MAEalpha=0.3, beta=0.72023.06 | 78.8 | — | — | — | — | — | — | — | — | |
| T-Lossalpha=0.7, beta=0.52023.06 | 78.8 | — | — | — | — | — | — | — | — | |
| MAEalpha=0.3, beta=0.72023.06 | 78.6 | — | — | — | — | — | — | — | — | |
| GCEalpha=0.3, beta=0.72023.06 | 78.5 | — | — | — | — | — | — | — | — | |
| RCEalpha=0.5, beta=0.52023.06 | 77.9 | — | — | — | — | — | — | — | — | |
| NCE+RCEalpha=0.5, beta=0.52023.06 | 77.7 | — | — | — | — | — | — | — | — | |
| NGCE+RCEalpha=0.5, beta=0.52023.06 | 77.6 | — | — | — | — | — | — | — | — | |
| SCEalpha=0.5, beta=0.52023.06 | 77.4 | — | — | — | — | — | — | — | — | |
| NGCE+MAEalpha=0.5, beta=0.52023.06 | 77.4 | — | — | — | — | — | — | — | — | |
| NGCEalpha=0.5, beta=0.52023.06 | 77.3 | — | — | — | — | — | — | — | — | |
| GCEalpha=0.5, beta=0.52023.06 | 77.2 | — | — | — | — | — | — | — | — | |
| MAEalpha=0.5, beta=0.52023.06 | 77.1 | — | — | — | — | — | — | — | — | |
| T-Lossalpha=0.7, beta=0.72023.06 | 76.1 | — | — | — | — | — | — | — | — | |
| RPI-NetLearning Paradigm=Unsupervised2026.04 | 75.84 | — | 84.15 | 91.73 | 87.62 | 66.21 | — | — | — |