Semantic Segmentation on PASCAL VOC blended high-quality 1/16 partition 2012
80.9mIoUUniMatch
Evaluation Results
| Method | Links | |
|---|---|---|
| UniMatchBackbone=ResNet-101, Segmentation Framework=DeepLabv3+2023.10 | 80.9 | |
| Dual TeacherBackbone=ResNet-101, Segmentation Framework=DeepLabv3+, CutMix Variant=Original-CutMix2023.10 | 80.07 | |
| AugSegBackbone=ResNet-101, Segmentation Framework=DeepLabv3+2023.10 | 79.29 | |
| Dual TeacherBackbone=ResNet-101, Segmentation Framework=DeepLabv3+, CutMix Variant=Adaptive-CutMix2023.10 | 78.82 | |
| Dual TeacherBackbone=ResNet-50, Segmentation Framework=DeepLabv3+, CutMix Variant=Original-CutMix2023.10 | 78.66 | |
| UniMatchBackbone=ResNet-50, Segmentation Framework=DeepLabv3+2023.10 | 78.1 | |
| AugSegBackbone=ResNet-50, Segmentation Framework=DeepLabv3+2023.10 | 77.28 | |
| U2PLBackbone=ResNet-101, Segmentation Framework=DeepLabv3+2023.10 | 77.21 | |
| Dual TeacherBackbone=ResNet-50, Segmentation Framework=DeepLabv3+, CutMix Variant=Adaptive-CutMix2023.10 | 77 | |
| U2PLBackbone=ResNet-50, Segmentation Framework=DeepLabv3+2023.10 | 74.7 | |
| Supervised-onlyBackbone=ResNet-101, Segmentation Framework=DeepLabv3+2023.10 | 70.6 | |
| Supervised-onlyBackbone=ResNet-50, Segmentation Framework=DeepLabv3+2023.10 | 67.7 |