Semantic Segmentation on PascalVOC 12 (val)
62.09mIoU OverallAlignSAM
Evaluation results
| Method | Links | ||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AlignSAM2024.06 | 62.09 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 48.09 | 66.13 | 73.12 | 85.99 | 31.68 | 79.84 | 64.72 | 61.63 | 72.97 | 75.1 | |
| Painter2024.06 | 59.27 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 35.33 | 61.93 | 77.29 | 63.16 | 29.9 | 59.82 | 46.4 | 55.58 | 78.71 | 72.82 | |
| AlignSAMreinforcement learning=false2024.06 | 54.06 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 41.69 | 56.28 | 69.3 | 80.8 | 26.84 | 69.16 | 50.83 | 48.2 | 62.74 | 72.91 | |
| PerSAM2024.06 | 53.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36.03 | 46.94 | 64.42 | 69.39 | 22.28 | 67.25 | 49.07 | 36.84 | 68.79 | 65.56 | |
| SEEM2024.06 | 52.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36.95 | 44.95 | 63.93 | 84.46 | 19.67 | 66.4 | 63.87 | 47.89 | 73.74 | 74.1 | |
| SAMed2024.06 | 51.42 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 36.13 | 48.72 | 60.59 | 58.46 | 16.22 | 71.48 | 49.08 | 47.65 | 74.33 | 71.72 | |
| MSA2024.06 | 47.12 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 31.56 | 41.05 | 56.47 | 60.65 | 19.79 | 55.95 | 38.81 | 33.97 | 60.38 | 61.99 | |
| FBLFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 43.9 | — | — | — | 86.3 | 66.2 | 34 | 48.3 | 28 | 6.9 | 64.7 | 75.6 | 74.1 | 0 | 26 | 29.9 | 61.7 | 40.1 | 6,600 | 70.4 | 0 | 40.4 | 27.4 | 26.8 | 48.5 | — | — | — | — | — | — | — | — | — | — | — | |
| MiBFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 33 | — | — | — | 86.7 | 50.9 | 23 | 17.7 | 25 | 8.9 | 41.3 | 67.1 | 47.9 | 4.5 | 0.1 | 29.1 | 26.9 | 21.7 | 6,980 | 73.2 | 3.1 | 17.9 | 30.3 | 29.2 | 19.8 | 10.9 | — | — | — | — | — | — | — | — | — | — | |
| RCILFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 32.4 | — | — | — | 85.6 | 62.8 | 29.6 | 38.9 | 39.3 | 0.9 | 62.3 | 51.2 | 32.6 | 0.3 | 34.1 | 21.1 | 3.9 | 18.1 | 4,080 | 68.6 | 1.2 | 6.5 | 27.7 | 15 | 39.1 | 11.5 | — | — | — | — | — | — | — | — | — | — | |
| PLOPFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 28.1 | — | — | — | 85.5 | 1.7 | 0.3 | 0 | 44.3 | 0.2 | 66.1 | 58.1 | 0.6 | 0 | 1.9 | 25.1 | 33.4 | 31 | 4,610 | 70.3 | 0 | 27.5 | 25 | 36.1 | 36.5 | 15.8 | — | — | — | — | — | — | — | — | — | — | |
| AlignSAMsemantic recalibration module=false2024.06 | 27.73 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 19.78 | 27.42 | 30.33 | 39.29 | 14.08 | 29.16 | 24.91 | 30.39 | 31.32 | 29.02 | |
| LWFFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 23.8 | — | — | — | 88.4 | 0 | 0 | 0 | 0.9 | 0 | 0 | 8.7 | 16.3 | 8.5 | 0 | 39.2 | 39.6 | 38.6 | 6,320 | 77.7 | 24.9 | 15.1 | 24.9 | 25.1 | 29.6 | 20.1 | — | — | — | — | — | — | — | — | — | — | |
| ILTFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 22.7 | — | — | — | 87.8 | 0 | 0 | 0 | 8.5 | 0.1 | 4.1 | 22.7 | 14.5 | 2.4 | 0 | 34.5 | 25.2 | 36 | 6,350 | 74.4 | 15.2 | 13.5 | 23.4 | 24.7 | 26 | 21.2 | — | — | — | — | — | — | — | — | — | — | |
| FinetuningFramework=Federated Learning (FL), Setting=4-4 overlapped2023.04 | 9.1 | — | — | — | 73.1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 50 | 0 | 0 | 20.1 | 34.4 | 32.3 | 30.4 | 34.8 | — | — | — | — | — | — | — | — | — | — | |
| CaGNetBackbone=ResNet-101, Segmentation Model=DeepLabV22021.04 | — | 78.4 | 25.6 | 39.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CaGNet + STBackbone=ResNet-101, Segmentation Model=DeepLabV2, Self-training=True2021.04 | — | 78.6 | 30.3 | 43.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SPNetBackbone=ResNet-101, Segmentation Model=DeepLabV22021.04 | — | 73.3 | 15 | 21.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SPNet+STBackbone=ResNet-101, Segmentation Model=DeepLabV2, Self-training=Hard pseudo-labelling without consistency2021.04 | — | 77.8 | 25.8 | 38.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| STRICTBackbone=ResNet-101, Segmentation Model=DeepLabV2, Self-training=Consistency regularization2021.04 | — | 82.7 | 35.6 | 49.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ZS3Backbone=ResNet-101, Segmentation Model=DeepLabV22021.04 | — | 77.3 | 17.7 | 28.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ZS5Backbone=ResNet-101, Segmentation Model=DeepLabV2, Self-training=True2021.04 | — | 78 | 21.2 | 33.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |