Semantic Segmentation on ScanNet 20 Classes (test)
62.6mIoUDVEFormer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DVEFormerBackbone=SwinV2-T-128, Extra Training Data=true, Evaluation Protocol=Linear probing2026.01 | 62.6 | — | |
| CMXBackbone=2x MiT-B22026.01 | 61.3 | — | |
| EMSANetBackbone=2x ResNet34-NBt1D, Extra Training Data=true2026.01 | 60 | — | |
| EMSAFormerBackbone=SwinV2-T-1282026.01 | 56.4 | — |