Phase Recognition on Cholec80 (test)
0.874F1 ScoreMTRCNet-CL
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MTRCNet-CLmulti-task learning=true2019.07 | 0.874 | 0.892 | 0.869 | 0.88 | — | — | |
| EndoNet+LSTMmulti-task learning=true2019.07 | 0.845 | 0.886 | 0.844 | 0.847 | — | — | |
| SV-RCNetmulti-task learning=false2019.07 | 0.821 | 0.853 | 0.807 | 0.835 | — | — | |
| ImageNet SupervisedBackbone=ViT-B, Number of Labeled Videos=402023.12 | 0.804 | — | — | — | — | — | |
| ImageNet SupervisedBackbone=ResNet-50, Number of Labeled Videos=402023.12 | 0.803 | — | — | — | — | — | |
| AutoLaparo SSLBackbone=ResNet-50, Number of Labeled Videos=402023.12 | 0.8 | — | — | — | — | — | |
| Cholec80 SSLBackbone=ResNet-50, Number of Labeled Videos=402023.12 | 0.796 | — | — | — | — | — | |
| EndoFMBackbone=ViT-B, Number of Labeled Videos=402023.12 | 0.784 | — | — | — | — | — | |
| EndoNetmulti-task learning=true2019.07 | 0.765 | 0.817 | 0.737 | 0.796 | — | — | |
| PhaseNetmulti-task learning=false2019.07 | 0.738 | 0.788 | 0.713 | 0.766 | — | — | |
| Cholec80 SSLBackbone=ResNet-50, Number of Labeled Videos=52023.12 | 0.661 | — | — | — | — | — | |
| EndoFMBackbone=ViT-B, Number of Labeled Videos=52023.12 | 0.659 | — | — | — | — | — | |
| AutoLaparo SSLBackbone=ResNet-50, Number of Labeled Videos=52023.12 | 0.624 | — | — | — | — | — | |
| ImageNet SupervisedBackbone=ResNet-50, Number of Labeled Videos=52023.12 | 0.623 | — | — | — | — | — | |
| Cholec80 SSLBackbone=ResNet-50, Number of Labeled Videos=32023.12 | 0.617 | — | — | — | — | — | |
| ImageNet SupervisedBackbone=ViT-B, Number of Labeled Videos=52023.12 | 0.601 | — | — | — | — | — | |
| EndoFMBackbone=ViT-B, Number of Labeled Videos=32023.12 | 0.595 | — | — | — | — | — | |
| Binary Toolmulti-task learning=false2019.07 | 0.572 | 0.475 | 0.544 | 0.602 | — | — | |
| AutoLaparo SSLBackbone=ResNet-50, Number of Labeled Videos=32023.12 | 0.544 | — | — | — | — | — | |
| ImageNet SupervisedBackbone=ViT-B, Number of Labeled Videos=32023.12 | 0.543 | — | — | — | — | — | |
| ImageNet SupervisedBackbone=ResNet-50, Number of Labeled Videos=32023.12 | 0.504 | — | — | — | — | — | |
| Handcrafted + CCAmulti-task learning=false2019.07 | 0.404 | 0.382 | 0.394 | 0.415 | — | — | |
| HyperVLPPre-training Dataset=SVL, Zero-shot=true2026.06 | 0.372 | 49.9 | — | — | — | — | |
| PeskaVLPPre-training Dataset=SVL, Evaluation Protocol=Zero-shot2024.09 | 0.342 | 45.1 | — | — | — | — | |
| PeskaVLPPre-training Dataset=SVL, Zero-shot=true2026.06 | 0.342 | 45.1 | — | — | — | — | |
| HecVLPre-training Dataset=SVL, Evaluation Protocol=Zero-shot2024.09 | 0.263 | 41.7 | — | — | — | — | |
| HecVLPre-training Dataset=SVL, Zero-shot=true2026.06 | 0.263 | 41.7 | — | — | — | — | |
| SurgVLPPre-training Dataset=SVL, Evaluation Protocol=Zero-shot2024.09 | 0.244 | 34.7 | — | — | — | — | |
| SurgVLPPre-training Dataset=SVL, Zero-shot=true2026.06 | 0.244 | 34.7 | — | — | — | — | |
| CLIPPre-training Dataset=SVL, Evaluation Protocol=Zero-shot2024.09 | 0.196 | 33.8 | — | — | — | — | |
| CLIPPre-training Dataset=SVL, Zero-shot=true2026.06 | 0.196 | 33.8 | — | — | — | — | |
| CLIPPre-training Dataset=CLIP400M, Evaluation Protocol=Zero-shot2024.09 | 0.131 | 30.8 | — | — | — | — | |
| CLIPPre-training Dataset=CLIP400M, Zero-shot=true2026.06 | 0.131 | 30.8 | — | — | — | — | |
| CLIPPre-training Dataset=Scratch, Evaluation Protocol=Zero-shot2024.09 | 0.104 | 29.4 | — | — | — | — | |
| CLIPPre-training Dataset=Scratch, Zero-shot=true2026.06 | 0.104 | 29.4 | — | — | — | — | |
| MIL-NCEPre-training Dataset=Howto100M, Evaluation Protocol=Zero-shot2024.09 | 0.073 | 7.8 | — | — | — | — | |
| MIL-NCEPre-training Dataset=HowTo100M, Zero-shot=true2026.06 | 0.073 | 7.8 | — | — | — | — | |
| DINOArch=ViT-S, Pretrain=IN1K2023.08 | — | — | — | — | — | 64.9 | |
| DINOArch=RN50, Pretrain=Pub2023.08 | — | — | — | — | — | 71.1 | |
| EndoLSTM2020.03 | — | 80.85 | 76.81 | 72.07 | — | — | |
| FSArch=RN50, Pretrain=IN1K2023.08 | — | — | — | — | — | 71.5 | |
| MSNArch=ViT-S, Pretrain=Pub2023.08 | — | — | — | — | — | 65 | |
| MSNArch=ViT-B, Pretrain=Pub2023.08 | — | — | — | — | — | 71.2 | |
| MSNArch=ViT-L, Pretrain=Pub2023.08 | — | — | — | — | — | 65.6 | |
| MSNArch=ViT-S, Pretrain=Pri2023.08 | — | — | — | — | — | 70.7 | |
| MSNArch=ViT-B, Pretrain=Pri2023.08 | — | — | — | — | — | 73.5 | |
| MSNArch=ViT-L, Pretrain=Pri2023.08 | — | — | — | — | — | 76.3 | |
| MTRCNet2020.03 | — | 82.76 | 76.08 | 78.02 | — | — | |
| PhaseLSTM2020.03 | — | 79.68 | 72.85 | 73.45 | — | — | |
| ResNetLSTM2020.03 | — | 86.58 | 80.53 | 79.94 | — | — | |
| TeCNOfeature extractor=ResNet502020.03 | — | 88.56 | 81.64 | 85.24 | — | — |