Material Classification on GelSight sensor dataset Touch and Go (test)
59Material AccTaRF (Real + Estimated)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TaRF (Real + Estimated)Pre-training Dataset=Touch and Go + TaRF (Real + Estimated), Evaluation Protocol=Linear Probing2024.05 | 59 | 88.7 | 86.1 | |
| TaRF (Real)Pre-training Dataset=Touch and Go + TaRF (Real), Evaluation Protocol=Linear Probing2024.05 | 57.6 | 88.4 | 81.7 | |
| Touch and GoPre-training Dataset=Touch and Go, Evaluation Protocol=Linear Probing2024.05 | 54.7 | 77.3 | 79.4 | |
| Touch and Go + ObjectFolder 2.0Pre-training Dataset=Touch and Go + ObjectFolder 2.0, Evaluation Protocol=Linear Probing2024.05 | 54.6 | 87.3 | 84.8 | |
| Touch and Go + VisGelPre-training Dataset=Touch and Go + VisGel, Evaluation Protocol=Linear Probing2024.05 | 53.1 | 86.7 | 83.6 | |
| VisGelPre-training Dataset=VisGel, Evaluation Protocol=Linear Probing2024.05 | 39.1 | 69.4 | 70.4 | |
| ObjectFolder 2.0Pre-training Dataset=ObjectFolder 2.0, Evaluation Protocol=Linear Probing2024.05 | 36.2 | 72 | 69 | |
| ChancePre-training Dataset=None (Random), Evaluation Protocol=Linear Probing2024.05 | 18.6 | 66.1 | 56.3 |