Semantic Segmentation on Kvasir-SEG (test)
810.34IoUResUnet++
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| ResUnet++Flops/G=70.99, Params/M=14.482025.01 | 810.34 | — | — | — | — | — | — | |
| ResUnetFlops/G=80.98, Params/M=13.042025.01 | 800.14 | — | — | — | — | — | — | |
| SAMUSImage size=256 × 256, Random crop=224 × 224 + padding2026.05 | 90.13 | — | — | — | — | — | — | |
| GSAMImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 86.99 | — | — | — | — | — | — | |
| U-NetImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 86.58 | — | — | — | — | — | — | |
| HP-AdapterImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 85.84 | — | — | — | — | — | — | |
| LW-AdapterImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 84.75 | — | — | — | — | — | — | |
| CRISTraining Strategy=E2E fine-tuning, Trainable Parameters=147M2024.05 | 83.37 | 89.43 | 55.23 | — | — | — | — | |
| SAM2Image size=224 × 224, Random crop=224 × 224 + padding2026.05 | 82.99 | — | — | — | — | — | — | |
| CLIPSeg DATraining Strategy=Adapter fine-tuning, Trainable Parameters=3M2024.05 | 82.39 | 89.1 | 47.79 | — | — | — | — | |
| CLIPSegTraining Strategy=E2E fine-tuning, Trainable Parameters=150M2024.05 | 81.72 | 87.69 | 54.02 | — | — | — | — | |
| SM-AdaptFormerImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 80.48 | — | — | — | — | — | — | |
| SAMImage size=1024 × 1024, Random crop=None2026.05 | 79.36 | — | — | — | — | — | — | |
| CLIPSeg SATraining Strategy=Adapter fine-tuning, Trainable Parameters=4.2M2024.05 | 79.26 | 86.85 | 52.18 | — | — | — | — | |
| AdaptFormerImage size=224 × 224, Random crop=224 × 224 + padding2026.05 | 72.49 | — | — | — | — | — | — | |
| DistillFSSBackbone=Swin-B, Number of support images (K)=252025.12 | 61.87 | — | — | — | — | — | — | |
| TransferFSSBackbone=Swin-B, Number of support images (K)=252025.12 | 61.84 | — | — | — | — | — | — | |
| TransferFSSBackbone=Swin-B, Number of support images (K)=502025.12 | 59.97 | — | — | — | — | — | — | |
| SANTraining Strategy=Adapter fine-tuning, Trainable Parameters=8.4M2024.05 | 58.05 | 69.58 | 130.75 | — | — | — | — | |
| DistillFSSBackbone=Swin-B, Number of support images (K)=502025.12 | 57.09 | — | — | — | — | — | — | |
| DistillFSSBackbone=Swin-B, Number of support images (K)=102025.12 | 52.83 | — | — | — | — | — | — | |
| DMTNetBackbone=ResNet-50, Number of support images (K)=252025.12 | 50.01 | — | — | — | — | — | — | |
| PATNetBackbone=ResNet-50, Number of support images (K)=252025.12 | 49.72 | — | — | — | — | — | — | |
| TransferFSSBackbone=Swin-B, Number of support images (K)=102025.12 | 49.28 | — | — | — | — | — | — | |
| DMTNetBackbone=ResNet-50, Number of support images (K)=102025.12 | 48.98 | — | — | — | — | — | — | |
| TransferFSSBackbone=ResNet-50, Number of support images (K)=502025.12 | 48.58 | — | — | — | — | — | — | |
| DMTNetBackbone=ResNet-50, Number of support images (K)=52025.12 | 47.78 | — | — | — | — | — | — | |
| PATNetBackbone=ResNet-50, Number of support images (K)=502025.12 | 46.78 | — | — | — | — | — | — | |
| DMTNetBackbone=ResNet-50, Number of support images (K)=502025.12 | 46.54 | — | — | — | — | — | — | |
| DistillFSSBackbone=ResNet-50, Number of support images (K)=502025.12 | 46.18 | — | — | — | — | — | — | |
| TransferFSSBackbone=ResNet-50, Number of support images (K)=252025.12 | 45.72 | — | — | — | — | — | — | |
| TransferFSSBackbone=Swin-B, Number of support images (K)=52025.12 | 45.18 | — | — | — | — | — | — | |
| PATNetBackbone=ResNet-50, Number of support images (K)=102025.12 | 44.93 | — | — | — | — | — | — | |
| DistillFSSBackbone=ResNet-50, Number of support images (K)=252025.12 | 44.9 | — | — | — | — | — | — | |
| DistillFSSBackbone=ResNet-50, Number of support images (K)=102025.12 | 42.72 | — | — | — | — | — | — | |
| PATNetBackbone=ResNet-50, Number of support images (K)=52025.12 | 41.94 | — | — | — | — | — | — | |
| TransferFSSBackbone=ResNet-50, Number of support images (K)=102025.12 | 41.18 | — | — | — | — | — | — | |
| DistillFSSBackbone=Swin-B, Number of support images (K)=52025.12 | 37.29 | — | — | — | — | — | — | |
| DCAMABackbone=ResNet-50, Number of support images (K)=52025.12 | 36.08 | — | — | — | — | — | — | |
| DCAMABackbone=ResNet-50, Number of support images (K)=252025.12 | 34.34 | — | — | — | — | — | — | |
| HDMNetBackbone=ResNet-50, Number of support images (K)=252025.12 | 34.13 | — | — | — | — | — | — | |
| DCAMABackbone=ResNet-50, Number of support images (K)=102025.12 | 33.4 | — | — | — | — | — | — | |
| HDMNetBackbone=ResNet-50, Number of support images (K)=502025.12 | 33.37 | — | — | — | — | — | — | |
| TransferFSSBackbone=ResNet-50, Number of support images (K)=52025.12 | 31.99 | — | — | — | — | — | — | |
| DCAMABackbone=ResNet-50, Number of support images (K)=502025.12 | 29.82 | — | — | — | — | — | — | |
| DistillFSSBackbone=ResNet-50, Number of support images (K)=52025.12 | 29.75 | — | — | — | — | — | — | |
| HDMNetBackbone=ResNet-50, Number of support images (K)=102025.12 | 29.7 | — | — | — | — | — | — | |
| DCAMABackbone=Swin-B, Number of support images (K)=52025.12 | 29.42 | — | — | — | — | — | — | |
| HDMNetBackbone=ResNet-50, Number of support images (K)=52025.12 | 28.97 | — | — | — | — | — | — | |
| DCAMABackbone=Swin-B, Number of support images (K)=102025.12 | 28.65 | — | — | — | — | — | — | |
| DCAMABackbone=Swin-B, Number of support images (K)=252025.12 | 28.37 | — | — | — | — | — | — | |
| DCAMABackbone=Swin-B, Number of support images (K)=502025.12 | 28.1 | — | — | — | — | — | — | |
| LabelAnythingBackbone=ViT-B, Number of support images (K)=102025.12 | 27.78 | — | — | — | — | — | — | |
| LabelAnythingBackbone=ViT-B, Number of support images (K)=52025.12 | 27.76 | — | — | — | — | — | — | |
| LabelAnythingBackbone=ViT-B, Number of support images (K)=252025.12 | 27.75 | — | — | — | — | — | — | |
| LabelAnythingBackbone=ViT-B, Number of support images (K)=502025.12 | 27.67 | — | — | — | — | — | — | |
| BAMBackbone=ResNet-50, Number of support images (K)=252025.12 | 23.05 | — | — | — | — | — | — | |
| BAMBackbone=ResNet-50, Number of support images (K)=502025.12 | 23.03 | — | — | — | — | — | — | |
| BAMBackbone=ResNet-50, Number of support images (K)=102025.12 | 19.28 | — | — | — | — | — | — | |
| BAMBackbone=ResNet-50, Number of support images (K)=52025.12 | 18.96 | — | — | — | — | — | — | |
| FCNFlops/G=34.71, Params/M=32.942025.01 | — | — | — | — | 86.621 | — | — | |
| UCTransNetFlops/G=43.06, Params/M=66.242025.01 | — | — | — | — | — | 88.331 | — |