Object Category Classification on Touch3D
73.51Top-1 AccuracyPatch Vis + CL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Patch Vis + CLmodality=Vision, input_type=Patch, alignment=CL2026.06 | 73.51 | 92.59 | |
| Patch Vis + CLmodality=Vision, input_type=Patch, alignment=CL2026.06 | 70.4 | — | |
| Patch Vis + DARmodality=Vision, input_type=Patch, alignment=DAR2026.06 | 69.78 | 87.29 | |
| Patch VisTac Early Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Early2026.06 | 69.78 | 87.29 | |
| Patch Vis + CLmodality=Vision, input_type=Patch, alignment=CL2026.06 | 69.5 | 89.44 | |
| Patch Tac + DARmodality=Tactile, input_type=Patch, alignment=DAR2026.06 | 67.67 | — | |
| Patch Vismodality=Vision, input_type=Patch2026.06 | 67.38 | 85.96 | |
| Patch Vis + DARmodality=Vision, input_type=Patch, alignment=DAR2026.06 | 63.82 | — | |
| Patch VisTac Early Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Early2026.06 | 62.62 | 85.18 | |
| Patch Vismodality=Vision, input_type=Patch2026.06 | 60.32 | — | |
| Patch VisTac Early Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Early2026.06 | 60.06 | — | |
| Vismodality=Vision2026.06 | 59.04 | — | |
| Patch Tac + DARmodality=Tactile, input_type=Patch, alignment=DAR2026.06 | 58.68 | 79.98 | |
| Patch Tac + CLmodality=Tactile, input_type=Patch, alignment=CL2026.06 | 58.64 | — | |
| Patch VisTac Early Fusionmodality=Vision-Tactile, input_type=Patch, fusion=Early2026.06 | 58.32 | 79.37 | |
| Patch VisTac Late Fusionmodality=Vision-Tactile, input_type=Patch, fusion=Late2026.06 | 58.05 | — | |
| Patch Vis + DARmodality=Vision, input_type=Patch, alignment=DAR2026.06 | 57.32 | 78.44 | |
| Patch VisTac Late Fusionmodality=Vision-Tactile, input_type=Patch, fusion=Late2026.06 | 56.99 | 79.22 | |
| Patch VisTac Early Fusionmodality=Vision-Tactile, input_type=Patch, fusion=Early2026.06 | 56.13 | — | |
| Patch VisTac Early Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Early2026.06 | 55.96 | 79.09 | |
| Patch Vis + DSCMRmodality=Vision, input_type=Patch, alignment=DSCMR2026.06 | 55.86 | — | |
| Vismodality=Vision2026.06 | 54.74 | 76.58 | |
| Patch VisTac Early Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Early2026.06 | 54.71 | — | |
| Patch Tacmodality=Tactile, input_type=Patch2026.06 | 54.18 | — | |
| Patch VisTac Late Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Late2026.06 | 54.18 | — | |
| Patch Tac + DSCMRmodality=Tactile, input_type=Patch, alignment=DSCMR2026.06 | 54.01 | — | |
| Patch Vismodality=Vision, input_type=Patch, alignment=None2026.06 | 52.63 | 74.69 | |
| Patch VisTac Early Fusionmodality=Vision-Tactile, input_type=Patch, alignment=None, fusion=Early2026.06 | 51.69 | 72 | |
| Tac + GAmodality=Tactile, alignment=GA2026.06 | 51.44 | — | |
| Patch Tac + CLmodality=Tactile, input_type=Patch, alignment=CL2026.06 | 50.91 | 70.96 | |
| Patch VisTac Early Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Early2026.06 | 50.84 | 73.83 | |
| Patch VisTac Late Fusionmodality=Vision-Tactile, input_type=Patch, alignment=None, fusion=Late2026.06 | 50.43 | 71.03 | |
| Vismodality=Vision, input_type=Normal, alignment=None2026.06 | 49.12 | 70.75 | |
| Patch VisTac Late Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Late2026.06 | 49.05 | 73.79 | |
| Patch VisTac Late Fusion+DARmodality=Vision-Tactile, input_type=Patch, alignment=DAR, fusion=Late2026.06 | 48.51 | 73.79 | |
| Tacmodality=Tactile2026.06 | 48.38 | — | |
| Patch Tac + DARmodality=Tactile, input_type=Patch, alignment=DAR2026.06 | 48.3 | 70.35 | |
| Patch Vis + DSCMRmodality=Vision, input_type=Patch, alignment=DSCMR2026.06 | 47.94 | 76.69 | |
| Patch Vis + DSCMRmodality=Vision, input_type=Patch, alignment=DSCMR2026.06 | 41.42 | 71.25 | |
| Vis + GAmodality=Vision, input_type=Normal, alignment=GA2026.06 | 41.21 | 68.99 | |
| Patch VisTac Late Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Late2026.06 | 41.1 | 64.02 | |
| Patch VisTac Late Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Late2026.06 | 40.28 | — | |
| Vis + GAmodality=Vision, alignment=GA2026.06 | 40.21 | 65.02 | |
| Patch VisTac Late Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Late2026.06 | 39.13 | 67.06 | |
| Patch VisTac Late Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Late2026.06 | 38.87 | — | |
| Patch VisTac Early Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Early2026.06 | 37.38 | 63.52 | |
| Patch VisTac Early Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Early2026.06 | 36.23 | — | |
| Patch VisTac Early Fusion+CLmodality=Vision-Tactile, input_type=Patch, alignment=CL, fusion=Early2026.06 | 35.12 | 67.78 | |
| Patch Tac + CLmodality=Tactile, input_type=Patch, alignment=CL2026.06 | 34.23 | 56.39 | |
| Patch VisTac Late Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Late2026.06 | 33.08 | 66.3 | |
| Patch VisTac Late Fusion+DSCMRmodality=Vision-Tactile, input_type=Patch, alignment=DSCMR, fusion=Late2026.06 | 32.65 | 58.83 | |
| Patch Tac + DSCMRmodality=Tactile, input_type=Patch, alignment=DSCMR2026.06 | 30.18 | 56.03 | |
| Patch Tac + DSCMRmodality=Tactile, input_type=Patch, alignment=DSCMR2026.06 | 30.15 | 55.1 | |
| Vis + GAmodality=Vision, alignment=GA2026.06 | 28.07 | — | |
| Patch Tacmodality=Tactile, input_type=Patch2026.06 | 23.17 | 49.95 | |
| Patch Tacmodality=Tactile, input_type=Patch, alignment=None2026.06 | 21.05 | 49.34 | |
| Tacmodality=Tactile, input_type=Normal, alignment=None2026.06 | 14.75 | 40.17 | |
| Tac + GAmodality=Tactile, input_type=Normal, alignment=GA2026.06 | 13.97 | 37.01 | |
| Tac + GAmodality=Tactile, alignment=GA2026.06 | 12.82 | 26.24 | |
| Tacmodality=Tactile2026.06 | 8.95 | 31.61 | |
| Chance2026.06 | 6.17 | — | |
| Chanceinput_type=random baseline2026.06 | 3.01 | 16.71 | |
| Chance2026.06 | 2.37 | 15.44 |