Action Recognition on Toyota SmartHome (TSH) (CV1)
55.6Accuracyπ-ViT + 3D Poses
Evaluation results
| Method | Links | |
|---|---|---|
| π-ViT + 3D PosesPose=Yes, RGB=Yes2023.11 | 55.6 | |
| π-ViTModality - 2D Ske=training only, Modality - 3D Ske=true, Modality - RGB=true, Modeling Paradigm=Multimodal (separate modeling), Fusion=late fusion2025.12 | 55.6 | |
| π-ViT + 3D PosesPose=Yes, RGB=Yes2026.03 | 55.6 | |
| π-ViTPose=Only during training, RGB=Yes2023.11 | 55.2 | |
| π-ViTModality - 2D Ske=training only, Modality - 3D Ske=training only, Modality - RGB=true, Modeling Paradigm=Multimodal (separate modeling)2025.12 | 55.2 | |
| π-ViTPose=Training Only, RGB=Yes2026.03 | 55.2 | |
| TimeSformer + 2D-SIMPose=Only during training, RGB=Yes2023.11 | 54.8 | |
| TimeSformer + 3D-SIMPose=Only during training, RGB=Yes2023.11 | 51.2 | |
| PAN (Even)Modality - 2D Ske=false, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Unimodal, Sampling Strategy=zero padding2025.12 | 50.2 | |
| TimeSformerPose=No, RGB=Yes2023.11 | 50 | |
| TimeSformerModality - 2D Ske=false, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Unimodal2025.12 | 50 | |
| PAN-EnsembleModality - 2D Ske=false, Modality - 3D Ske=true, Modality - RGB=true, Modeling Paradigm=Multimodal (separate modeling)2025.12 | 49.2 | |
| PAN-UnifiedModality - 2D Ske=false, Modality - 3D Ske=true, Modality - RGB=true, Modeling Paradigm=Multimodal (unified modeling)2025.12 | 48.2 | |
| PAN (Even)Modality - 2D Ske=false, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Unimodal, Sampling Strategy=token replication2025.12 | 47.9 | |
| PAN (Guided)Modality - 2D Ske=true, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Multimodal (separate modeling)2025.12 | 46.5 | |
| SV-data2vecPose=Training Only, RGB=Yes2026.03 | 46 | |
| MotionFormerPose=No, RGB=Yes2023.11 | 45.2 | |
| MotionFormerModality - 2D Ske=false, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Unimodal2025.12 | 45.2 | |
| Proposed Multi-modal ArchitecturePose=Yes, RGB=Yes2026.03 | 44.2 | |
| VPNRGB=true, Pose=false, Pre-training=Kinetics-4002021.07 | 43.8 | |
| VPNPose=Yes, RGB=Yes2023.11 | 43.8 | |
| VPNPose=Yes, RGB=Yes2026.03 | 43.8 | |
| DG-STGCNModality=2-stream (joint+bone), Pre-training=NTU120-XSet2022.10 | 41.8 | |
| Neural Projection Layer2021.03 | 39.6 | |
| NPLRGB=true, Pose=false, Pre-training=Kinetics-4002021.07 | 39.6 | |
| MMNetPose=Yes, RGB=Yes2026.03 | 37.4 | |
| Video SwinPose=No, RGB=Yes2023.11 | 36.6 | |
| Video SwinModality - 2D Ske=false, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Unimodal2025.12 | 36.6 | |
| UNIKRGB=false, Pose=true, Pre-training=Posetics(Ours)2021.07 | 36.1 | |
| UNIK++* w. PRSModality=2-stream (joint+bone), Pose Input Quality=high-quality 2D Pose, Pre-training=Posetics, w. PRS=true2022.10 | 36.1 | |
| LLM-AR2024.03 | 36.1 | |
| ViAPose=Yes, RGB=No2026.03 | 35.6 | |
| STA2021.03 | 35.2 | |
| Separable STARGB=true, Pose=false, Pre-training=Kinetics-4002021.07 | 35.2 | |
| Separable STAPose=Yes, RGB=Yes2023.11 | 35.2 | |
| Separable STAPose=Yes, RGB=Yes2026.03 | 35.2 | |
| P-I3DPose=Yes, RGB=Yes2023.11 | 35.1 | |
| P-I3DPose=Yes, RGB=Yes2026.03 | 35.1 | |
| I3D2021.03 | 34.9 | |
| I3DRGB=true, Pose=false, Pre-training=Kinetics-4002021.07 | 34.9 | |
| I3DPose=No, RGB=Yes2026.03 | 34.9 | |
| UNIK2024.03 | 33.4 | |
| 2S-AGCN2024.03 | 32.2 | |
| HyperformerPose=Yes, RGB=No2023.11 | 31.6 | |
| HyperformerModality - 2D Ske=false, Modality - 3D Ske=true, Modality - RGB=false, Modeling Paradigm=Unimodal2025.12 | 31.6 | |
| ML-STGNet2024.03 | 29.9 | |
| DG-STGCNModality=2-stream (joint+bone)2022.10 | 23.2 | |
| UNIKRGB=false, Pose=true, Pre-training=Scratch2021.07 | 22.9 | |
| UNIKPose=Yes, RGB=No2026.03 | 22.9 | |
| SSTA-PRS2024.03 | 22.8 | |
| 2s-AGCNRGB=false, Pose=true, Pre-training=Scratch2021.07 | 22.5 | |
| AGCN++ w. PRSModality=2-stream (joint+bone), Pose Input Quality=high-quality 2D Pose, w. PRS=true2022.10 | 22.5 | |
| AGCN++Modality=2-stream (joint+bone), Pose Input Quality=high-quality 2D Pose2022.10 | 22.1 | |
| 2s-AGCNPose=Yes, RGB=No2023.11 | 21.6 | |
| 2s-AGCNModality - 2D Ske=false, Modality - 3D Ske=true, Modality - RGB=false, Modeling Paradigm=Unimodal2025.12 | 21.6 | |
| 2s-AGCNPose=Yes, RGB=No2026.03 | 21.6 | |
| PoseC3DPose=Yes, RGB=Yes2023.11 | 21.5 | |
| PoseC3DModality - 2D Ske=re-estimated data, Modality - 3D Ske=false, Modality - RGB=true, Modeling Paradigm=Multimodal (separate modeling)2025.12 | 21.5 | |
| IDT2021.03 | 20.9 | |
| PoseC3DPose=Yes, RGB=No2023.11 | 20 | |
| PoseC3DModality - 2D Ske=re-estimated data, Modality - 3D Ske=false, Modality - RGB=false, Modeling Paradigm=Unimodal2025.12 | 20 | |
| MS-G3D NetRGB=false, Pose=true, Pre-training=Scratch2021.07 | 17.5 | |
| MS-G3D NetPose=Yes, RGB=No2026.03 | 17.5 | |
| ST-GCNRGB=false, Pose=true, Pre-training=Scratch2021.07 | 15.5 | |
| Pose LSTM2021.03 | 13.4 | |
| LSTMRGB=false, Pose=true, Pre-training=Scratch2021.07 | 13.1 | |
| LSTM2022.10 | 13.1 | |
| ST-GCNPose=Yes, RGB=No2026.03 | 12.5 |