Human Interaction Intent Prediction on HRI Intent-Prediction Offline (Cross-Validation)
0.951AUROCMINT-RVAE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MINT-RVAESensors=RGB-only, Temporal Resolution=Frame + Sequence-level, Features=Pose + Emotion, Data Re-balancing=Yes (MINT-RVAE), Deployment Platform=CPU-only (Raspberry Pi 5)2025.12 | 0.951 | — | — | |
| Arreghini et al.Sensors=RGB–D + gaze, Temporal Resolution=Sequence-level, Features=Pose, gaze cues, Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | 0.912 | — | — | |
| Abbate et al.Sensors=RGB–D, Temporal Resolution=Sequence-level, Features=Pose (self-supervised), Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | 0.9 | — | — | |
| Kedia et al.Sensors=RGB–D, Temporal Resolution=Sequence-level, Features=Pose, robot action, Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | — | 89 | — | |
| Liu et al.Sensors=RGB–D, Temporal Resolution=Sequence-level, Features=3D skeleton, Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | — | — | 0.84 | |
| Trick et al.Sensors=RGB–D + audio, Temporal Resolution=Sequence-level, Features=Pose, gaze, speech, Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | — | — | 0.81 | |
| Zimmermann et al.Sensors=RGB–D, Temporal Resolution=Sequence-level, Features=Pose, gesture, Data Re-balancing=No, Deployment Platform=GPU workstation2025.12 | — | 87 | — |