Semantic Segmentation on Pascal VOC high-quality (test)
88.9mIoUUniMatch V2 + ECOCSeg
Evaluation results
| Method | Links | |
|---|---|---|
| UniMatch V2 + ECOCSegLabeled data ratio=Full, Number of labeled images=1464, Backbone=DINOv2-S2025.12 | 88.9 | |
| UniMatch V2Labeled data ratio=Full, Number of labeled images=1464, Backbone=DINOv2-S2025.12 | 87.8 | |
| UniMatch V2 + ECOCSegLabeled data ratio=1/2, Number of labeled images=732, Backbone=DINOv2-S2025.12 | 87.8 | |
| UniMatch V2 + ECOCSegLabeled data ratio=1/4, Number of labeled images=366, Backbone=DINOv2-S2025.12 | 87.1 | |
| UniMatch V2Labeled data ratio=1/2, Number of labeled images=732, Backbone=DINOv2-S2025.12 | 86.7 | |
| UniMatch V2 + ECOCSegLabeled data ratio=1/8, Number of labeled images=183, Backbone=DINOv2-S2025.12 | 86.6 | |
| UniMatch V2Labeled data ratio=1/4, Number of labeled images=366, Backbone=DINOv2-S2025.12 | 85.9 | |
| UniMatch V2Labeled data ratio=1/8, Number of labeled images=183, Backbone=DINOv2-S2025.12 | 85.5 | |
| CorrMatchLabeled data ratio=Full, Number of labeled images=1464, Backbone=ResNet-1012025.12 | 81.8 | |
| BeyondPixelsLabeled data ratio=Full, Number of labeled images=1464, Backbone=ResNet-1012025.12 | 81.7 | |
| AugSegLabeled data ratio=Full, Number of labeled images=1464, Backbone=ResNet-1012025.12 | 81.4 | |
| UniMatch V2 + ECOCSegLabeled data ratio=1/16, Number of labeled images=92, Backbone=DINOv2-S2025.12 | 81.1 | |
| BeyondPixelsLabeled data ratio=1/2, Number of labeled images=732, Backbone=ResNet-1012025.12 | 80.8 | |
| CorrMatchLabeled data ratio=1/2, Number of labeled images=732, Backbone=ResNet-1012025.12 | 80.6 | |
| AugSegLabeled data ratio=1/2, Number of labeled images=732, Backbone=ResNet-1012025.12 | 80.3 | |
| BeyondPixelsLabeled data ratio=1/4, Number of labeled images=366, Backbone=ResNet-1012025.12 | 79.8 | |
| CorrMatchLabeled data ratio=1/4, Number of labeled images=366, Backbone=ResNet-1012025.12 | 79.4 | |
| UniMatch V2Labeled data ratio=1/16, Number of labeled images=92, Backbone=DINOv2-S2025.12 | 79 | |
| AugSegLabeled data ratio=1/4, Number of labeled images=366, Backbone=ResNet-1012025.12 | 78.8 | |
| BeyondPixelsLabeled data ratio=1/8, Number of labeled images=183, Backbone=ResNet-1012025.12 | 78.6 | |
| CorrMatchLabeled data ratio=1/8, Number of labeled images=183, Backbone=ResNet-1012025.12 | 78.5 | |
| BeyondPixelsLabeled data ratio=1/16, Number of labeled images=92, Backbone=ResNet-1012025.12 | 77.3 | |
| CorrMatchLabeled data ratio=1/16, Number of labeled images=92, Backbone=ResNet-1012025.12 | 76.4 | |
| AugSegLabeled data ratio=1/8, Number of labeled images=183, Backbone=ResNet-1012025.12 | 75.5 | |
| AugSegLabeled data ratio=1/16, Number of labeled images=92, Backbone=ResNet-1012025.12 | 71.1 |