Semantic Segmentation (Scribble-supervised) on PASCAL VOC 2012 (val)
78.1mIoUSeSAM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SeSAMDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 78.1 | 98.2 | |
| AGMM*DeepLab architecture=V3+, Backbone=ResNet-101, implementation=re-implementation2026.04 | 76.7 | 96.4 | |
| AGMMDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 76.4 | 96.1 | |
| TELDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 75.8 | 96.9 | |
| Pan et al.DeepLab architecture=V2, Backbone=ResNet-1012026.04 | 74.6 | 96 | |
| A2GNNDeepLab architecture=V2, Backbone=ResNet-1012026.04 | 74.3 | 95.5 | |
| SPMLDeepLab architecture=V2, Backbone=ResNet-1012026.04 | 74.2 | 95.5 | |
| DBFNetDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 72.5 | 91.2 | |
| FixmatchDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 71.3 | 89.6 | |
| SAM baselineDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 70.3 | 88.4 | |
| SupervisedDeepLab architecture=V3+, Backbone=ResNet-1012026.04 | 69.7 | 87.6 |