Semantic Segmentation on PASCAL VOC 2012 original (val)
80.47mIoUGTA-Seg
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GTA-SegNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 80.47 | — | — | — | — | |
| U2PLNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 79.49 | — | — | — | — | |
| ST++Number of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 79.12 | — | — | — | — | |
| AugSeg#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=512x5122023.10 | 78.82 | 64.22 | 72.17 | 76.17 | 77.4 | |
| UniMatch#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=321x3212023.10 | 78.7 | 71.9 | 72.5 | 76 | 77.4 | |
| GTA-SegNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 78.37 | — | — | — | — | |
| Dual Teacher#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=321x3212023.10 | 78.15 | 70.76 | 74.53 | 76.43 | 77.68 | |
| ST++Number of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 77.33 | — | — | — | — | |
| CutMixNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 76.54 | — | — | — | — | |
| U2PLNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 76.16 | — | — | — | — | |
| GTA-SegNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 75.57 | — | — | — | — | |
| ST++Number of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 74.59 | — | — | — | — | |
| PC2SegNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 74.15 | — | — | — | — | |
| CutMixNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 73.73 | — | — | — | — | |
| U2PLNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 73.66 | — | — | — | — | |
| PseudoSegNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 73.23 | — | — | — | — | |
| GTA-SegNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 73.16 | — | — | — | — | |
| PC2SegNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 73.05 | — | — | — | — | |
| Supervised-only#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=321x3212023.10 | 72.94 | 44.03 | 52.26 | 61.65 | 66.72 | |
| SupOnlyNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 72.5 | — | — | — | — | |
| PseudoSegNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 72.41 | — | — | — | — | |
| PC2Seg#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=512x5122023.10 | 72.26 | 56.9 | 64.63 | 67.62 | 70.9 | |
| SupOnlyNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 71.69 | — | — | — | — | |
| ST++Number of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 71.01 | — | — | — | — | |
| PseudoSeg#Params=43.6M, Backbone=ResNet-50, Architecture=DeepLabv3+, Training Resolution=512x5122023.10 | 71 | 54.89 | 61.88 | 64.85 | 70.42 | |
| MTNumber of labeled samples=1464, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 70.96 | — | — | — | — | |
| GTA-SegNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 70.02 | — | — | — | — | |
| PC2SegNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 69.78 | — | — | — | — | |
| MTNumber of labeled samples=732, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 69.51 | — | — | — | — | |
| CutMixNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 69.46 | — | — | — | — | |
| U2PLNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 69.15 | — | — | — | — | |
| PseudoSegNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 69.14 | — | — | — | — | |
| U2PLNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 67.98 | — | — | — | — | |
| PC2SegNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 66.28 | — | — | — | — | |
| SupOnlyNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 65.88 | — | — | — | — | |
| PseudoSegNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 65.5 | — | — | — | — | |
| ST++Number of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 65.23 | — | — | — | — | |
| MTNumber of labeled samples=366, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 63.86 | — | — | — | — | |
| CutMixNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 63.47 | — | — | — | — | |
| MTNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 58.93 | — | — | — | — | |
| PseudoSegNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 57.6 | — | — | — | — | |
| PC2SegNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 57 | — | — | — | — | |
| SupOnlyNumber of labeled samples=183, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 54.92 | — | — | — | — | |
| CutMixNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 52.16 | — | — | — | — | |
| MTNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 51.72 | — | — | — | — | |
| SupOnlyNumber of labeled samples=92, Backbone=ResNet-101, Decoder=DeepLabv3+2023.01 | 45.77 | — | — | — | — |