Polyp Segmentation on CVC-ClinicDB (test)
94.69DSCFCBFormer
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FCBFormer2022.08 | 94.69 | 90.2 | 94.41 | 95.25 | — | |
| PVT-GCASCADE2023.10 | 94.68 | 90.18 | — | — | — | |
| TGANetBackbone=ResNet502022.05 | 94.57 | 89.9 | 94.37 | 95.19 | 94.39 | |
| PVT-CASCADE2023.10 | 94.34 | 89.98 | — | — | — | |
| UGCANetNote=Averaged over 5 runs2023.07 | 94.3 | 89.6 | — | — | — | |
| MSRF-NetResolution=256x256, Split Type=Source data2021.11 | 94.2 | 90.43 | 95.67 | 94.27 | — | |
| MSRF-NetParameters=18.38M, FPS=12.52021.05 | 94.2 | 90.43 | 95.67 | 94.27 | — | |
| TransFuse-L*Variant=L*2022.05 | 94.2 | 89.7 | — | — | — | |
| MSRF-Net2022.08 | 94.2 | 90.43 | 95.67 | 94.27 | — | |
| TransFuse-L*Variant=Large2023.07 | 94.2 | 89.7 | — | — | — | |
| DeepLabV3+Backbone=ResNet502022.05 | 93.91 | 89.73 | 94.41 | 94.42 | 93.89 | |
| HarDNet-DFUSFPS=302022.09 | 93.9 | — | — | — | — | |
| Siamese-DiffusionBase Architecture=Polyp-PVT, Prior Type=Mask-only (M)2025.05 | 93.9 | 89.3 | — | — | — | |
| Siamese-DiffusionBase Architecture=CT-Net, Prior Type=Mask-only (M)2025.05 | 93.8 | 88.9 | — | — | — | |
| Polyp-PVTCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 93.72 | 88.97 | — | — | — | |
| Polyp-PVTBase Architecture=Polyp-PVT2025.05 | 93.7 | 88.9 | — | — | — | |
| ControlNetBase Architecture=Polyp-PVT, Prior Type=Mask-only (M)2025.05 | 93.7 | 89 | — | — | — | |
| T2I-AdapterBase Architecture=CT-Net, Prior Type=Mask-only (M)2025.05 | 93.7 | 88.7 | — | — | — | |
| CaraNet2022.05 | 93.6 | 88.7 | — | — | — | |
| CaraNet2023.07 | 93.6 | 88.7 | — | — | — | |
| CaraNet2023.10 | 93.6 | 88.7 | — | — | — | |
| CT-NetBase Architecture=CT-Net2025.05 | 93.6 | 88.7 | — | — | — | |
| SinGAN-SegBase Architecture=CT-Net, Prior Type=Mask-only (M)2025.05 | 93.6 | 88.8 | — | — | — | |
| TransUNet2022.05 | 93.5 | 88.7 | — | — | — | |
| TransUNet2023.07 | 93.5 | 88.7 | — | — | — | |
| Copy-PasteBase Architecture=Polyp-PVT2025.05 | 93.4 | 88.7 | — | — | — | |
| CFA-Net2024.08 | 93.3 | 88.3 | — | — | — | |
| GMSRF-NetResolution=256x256, Split Type=Source data2021.11 | 93.26 | 88.82 | 93.76 | 93.07 | — | |
| HarDNet-MSEGresolution=312x3122021.01 | 93.2 | 88.2 | — | — | — | |
| HarDNet-MSEG2022.05 | 93.2 | 88.2 | — | — | — | |
| ColonFormerVariant=S2022.05 | 93.2 | 88.3 | — | — | — | |
| ColonFormerVariant=L2022.05 | 93.2 | 88.4 | — | — | — | |
| HarDNet-MSEGFPS=1082022.09 | 93.2 | — | — | — | — | |
| HarDNet-MSEG2023.07 | 93.2 | 88.2 | — | — | — | |
| ColonFormer-SVariant=Small2023.07 | 93.2 | 83.3 | — | — | — | |
| ColonFormer-LVariant=Large2023.07 | 93.2 | 88.4 | — | — | — | |
| ControlNetBase Architecture=CT-Net, Prior Type=Mask-Image joint (M+I)2025.05 | 93.2 | 88.6 | — | — | — | |
| PraNetBackbone=Res2Net2022.05 | 93.18 | 88.66 | 93.47 | 94.79 | 93.33 | |
| T2I-AdapterBase Architecture=Polyp-PVT, Prior Type=Mask-only (M)2025.05 | 93.1 | 88.1 | — | — | — | |
| PolypPVT2023.10 | 93.08 | 88.28 | — | — | — | |
| Siamese-DiffusionBase Architecture=SANet, Prior Type=Mask-only (M)2025.05 | 93 | 88.1 | — | — | — | |
| ArSDMBase Architecture=CT-Net, Prior Type=Mask-only (M)2025.05 | 92.9 | 87.8 | — | — | — | |
| SSFormerPVT2023.10 | 92.88 | 88.27 | — | — | — | |
| DoubleU-NetP-values=1.000e+00*, Parameters=29.29M, FPS=7.462021.05 | 92.72 | 88.89 | 93.95 | 95.92 | — | |
| HRNetV2-W48Resolution=256x256, Split Type=Source data2021.11 | 92.44 | 87.47 | 92.34 | 92.96 | — | |
| HRNetV2-W48P-values=1.000e+00*, Parameters=65.84M, FPS=29.762021.05 | 92.44 | 87.47 | 92.34 | 92.96 | — | |
| DDANetP-values=1.000e+00*, Parameters=6.83M, FPS=58.152021.05 | 92.33 | 87.47 | 92.71 | 92.59 | — | |
| Copy-PasteBase Architecture=CT-Net2025.05 | 92.3 | 87.1 | — | — | — | |
| ControlNetBase Architecture=CT-Net, Prior Type=Mask-only (M)2025.05 | 92.3 | 87.5 | — | — | — | |
| ArSDMBase Architecture=Polyp-PVT, Prior Type=Mask-only (M)2025.05 | 92.2 | 87.5 | — | — | — | |
| ControlNetBase Architecture=Polyp-PVT, Prior Type=Mask-Image joint (M+I)2025.05 | 92.2 | 87.7 | — | — | — | |
| MSNetCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 92.19 | 87.98 | — | — | — | |
| SAM Adapter + SAIFCategory=SAM-based, Prompting Strategy=Box prompts, Inference Strategy=SAIF2026.03 | 92.17 | 85.22 | — | — | — | |
| CaraNet2024.08 | 92.1 | 87.6 | — | — | — | |
| UACANet-SResolution=256x256, Split Type=Source data2021.11 | 91.9 | 87 | 92.85 | 92.01 | — | |
| UACANet-SP-values=1.000e+00*, Parameters=26.90M, FPS=31.792021.05 | 91.9 | 87 | 92.85 | 92.01 | — | |
| AG-CUResNeSt-101Training scenario=Scenario 42021.05 | 91.7 | 86.7 | — | — | — | |
| SinGAN-SegBase Architecture=Polyp-PVT, Prior Type=Mask-only (M)2025.05 | 91.7 | 87 | — | — | — | |
| SANet2024.08 | 91.6 | 85.9 | — | — | — | |
| SANetBase Architecture=SANet2025.05 | 91.6 | 85.9 | — | — | — | |
| ControlNetBase Architecture=SANet, Prior Type=Mask-only (M)2025.05 | 91.6 | 85.4 | — | — | — | |
| U-NetResolution=256x256, Split Type=Source data2021.11 | 91.45 | 86.54 | 91.78 | 93.81 | — | |
| U-NetP-values=1.000e+00*, Parameters=7.11M, FPS=22.842021.05 | 91.45 | 86.54 | 91.78 | 93.81 | — | |
| ArSDMBase Architecture=SANet, Prior Type=Mask-only (M)2025.05 | 91.4 | 86.1 | — | — | — | |
| ControlNetBase Architecture=SANet, Prior Type=Mask-Image joint (M+I)2025.05 | 91.4 | 86.2 | — | — | — | |
| ColonSegNetResolution=256x256, Split Type=Source data2021.11 | 91.32 | 86 | 90.72 | 92.92 | — | |
| ColonSegNetP-values=1.000e+00*, Parameters=5.01M, FPS=7.982021.05 | 91.32 | 86 | 90.72 | 92.92 | — | |
| UACANet-L2023.10 | 91.07 | 86.7 | — | — | — | |
| UACANet-LResolution=256x256, Split Type=Source data2021.11 | 90.98 | 86.49 | 91.74 | 91.14 | — | |
| UACANet-LP-values=1.000e+00*, Parameters=69.15M, FPS=32.812021.05 | 90.98 | 86.49 | 91.74 | 91.14 | — | |
| SinGAN-SegBase Architecture=SANet, Prior Type=Mask-only (M)2025.05 | 90.9 | 85.3 | — | — | — | |
| ResUNet++P-values=1.000e+00*, Parameters=4.07M, FPS=15.712021.05 | 90.75 | 85.87 | 91.56 | 93.25 | — | |
| HRNetV2-W18-Smallv2Resolution=256x256, Split Type=Source data2021.11 | 90.73 | 84.57 | 91.37 | 91.91 | — | |
| HRNetV2-W18-Smallv2P-values=1.000e+00*, Parameters=26.20M, FPS=57.472021.05 | 90.73 | 84.57 | 91.37 | 91.91 | — | |
| PraNetResolution=256x256, Split Type=Source data2021.11 | 90.72 | 85.75 | 92.27 | 91.34 | — | |
| PraNetP-values=1.000e+00*, Parameters=32.54M, FPS=47.922021.05 | 90.72 | 85.75 | 92.27 | 91.34 | — | |
| HarDNet-MSEGTraining scenario=Scenario 4, Retrained with original reported configurations=true2021.05 | 90.7 | 85.3 | — | — | — | |
| SAM2-UNetResolution=352x352, Multi-scale training={1, 1.25}2024.08 | 90.7 | 85.6 | — | — | — | |
| T2I-AdapterBase Architecture=SANet, Prior Type=Mask-only (M)2025.05 | 90.6 | 85 | — | — | — | |
| SwinUNetCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 90.53 | 85.52 | — | — | — | |
| EU-NetCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 90.24 | 84.67 | — | — | — | |
| Copy-PasteBase Architecture=SANet2025.05 | 90.2 | 85.1 | — | — | — | |
| PraNetCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 90.2 | 85.82 | — | — | — | |
| PraNetresolution=312x3122021.01 | 89.9 | 84.9 | — | — | — | |
| PraNetTraining scenario=Scenario 42021.05 | 89.9 | 84.9 | — | — | — | |
| PraNet2022.05 | 89.9 | 84.9 | — | — | — | |
| PraNet2022.08 | 89.9 | 84.9 | — | — | — | |
| PraNet2023.07 | 89.9 | 84.9 | — | — | — | |
| PraNet2023.10 | 89.9 | 84.9 | — | — | — | |
| PraNet2024.08 | 89.9 | 84.9 | — | — | — | |
| Deeplabv3+Backbone=Mobilenet, Resolution=256x256, Split Type=Source data2021.11 | 89.85 | 85.88 | 91.6 | 92.87 | — | |
| Deeplabv3+Backbone=Mobilenet, P-values=2.838e-01, Parameters=2.14M, FPS=35.682021.05 | 89.85 | 85.88 | 91.6 | 92.87 | — | |
| U-Net2022.05 | 89.78 | 84.28 | 90.01 | 92.09 | 89.81 | |
| HarDNet-MSEGBackbone=HardNet682022.05 | 89.67 | 83.88 | 89.29 | 92.16 | 89.38 | |
| TransUNetCategory=Train-based, Prompting Strategy=Prompt-free2026.03 | 89.46 | 85.02 | — | — | — | |
| Deeplabv3+Backbone=Xception, Resolution=256x256, Split Type=Source data2021.11 | 88.97 | 87.06 | 92.51 | 93.66 | — | |
| Deeplabv3+Backbone=Xception, P-values=6.723e-01, Parameters=41.25M, FPS=29.082021.05 | 88.97 | 87.06 | 92.51 | 93.66 | — | |
| P2SAMAdaptation Strategy=Full-Fine-Tune, Backbone=large, Parameters Tuned=312.5M, Training-free=true, using SAM=true2024.03 | 88.76 | — | — | — | — | |
| ColonSegNet2022.05 | 88.62 | 82.48 | 88.28 | 90.17 | 88.26 | |
| ResUNet++ + CRFPost-processing=CRF, P-values=1.102e-01, Parameters=4.02M, FPS=8.552021.05 | 88.15 | 88.99 | 89.7 | 86.74 | — |