Semantic Segmentation on Stanford Indoor 1% labeled (491 samples)
49.28mIoUOurs (LF+M3L)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Ours (LF+M3L)Input Modality=RGB, Supervision=Semi-supervised (M3L)2023.04 | 49.28 | — | — | — | — | |
| Uni-modal + MTInput Modality=RGB, Supervision=Semi-supervised (Mean Teacher)2023.04 | 47.04 | — | — | — | — | |
| LF+M3LInput Modality=Depth, Training Strategy=Semi-supervised2023.04 | 46.79 | — | — | — | — | |
| Uni-modal (sup-only)Input Modality=RGB, Supervision=Supervised-only2023.04 | 46.45 | — | — | — | — | |
| Uni-modal + CPS-SegInput Modality=RGB, Supervision=Semi-supervised (CPS-Seg)2023.04 | 46.37 | — | — | — | — | |
| Uni-modal + MTInput Modality=Depth, Training Strategy=Semi-supervised2023.04 | 46.25 | — | — | — | — | |
| Uni-modal + CPS-SegInput Modality=Depth, Training Strategy=Semi-supervised2023.04 | 45.71 | — | — | — | — | |
| Uni-modal (sup-only)Input Modality=Depth, Training Strategy=Supervised-only2023.04 | 44.78 | — | — | — | — | |
| Linear Fusiontraining_mode=supervised2023.04 | — | 33.51 | 23.7 | 52.47 | 36.56 | |
| Linear Fusion + M3Ltraining_mode=semi-supervised2023.04 | — | 49.28 | 46.79 | 55.48 | 50.52 | |
| Linear Fusion + MTtraining_mode=semi-supervised2023.04 | — | 33.65 | 22.42 | 54.32 | 36.8 | |
| TFtraining_mode=supervised2023.04 | — | 37.34 | 28.33 | 51.85 | 39.17 | |
| Uni-modaltraining_mode=supervised2023.04 | — | 46.45 | 44.78 | — | — | |
| URNtraining_mode=supervised2023.04 | — | 36.25 | 33.27 | 52.07 | 40.53 |