Action Recognition on NTU RGB+D (Cross-View)
99AccuracyRNX3D101+MS-AAGCN-C
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RNX3D101+MS-AAGCN-CPose modality=true, RGB modality=true, Pose-guided cropping strategy=true2019.12 | 99 | — | — | — | — | — | — | — | — | — | — | — | — | |
| LLM-ARModality=Joint2024.03 | 98.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| KAN-HyperpointNet2026.06 | 98.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DVANetModality=RGB2023.12 | 98.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| RNX3D101+MS-AAGCNPose modality=true, RGB modality=true, Pose-guided cropping strategy=false2019.12 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMDR (CFCer)modality=RGB-D, #frames=642022.11 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shah et al.Modality=RGB2023.12 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SRENet2026.06 | 98 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMDR (CFCer)modality=RGB-D, #frames=322022.11 | 97.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SkateFormer (4 ensemble)Mod.=All, Param. (M)=2.03, FLOPs (G)=14.48, Architecture paradigm=ANN-Trans.2026.03 | 97.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SkeleTRModality=Joint2024.03 | 97.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PRG-Net2026.06 | 97.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SequentialPointNet2026.06 | 97.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PRENet2026.06 | 97.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMDR (CFCer)modality=RGB-D, #frames=162022.11 | 97.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DG-STGCNModality=Skeleton2023.12 | 97.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ST-GCN++Training Mode=With good practices and spatial augmentations2022.05 | 97.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Duan et al.Modality=Skeleton2023.12 | 97.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DTN(CAPF)Modality=RGB+Depth2021.12 | 97.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CAPFModality=RGB+Depth2021.12 | 97.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Decouple+Recouplemodality=RGB-D, #frames=642022.11 | 97.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Action MachinePose=false, RGB=true2018.12 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SDS-CLEvaluation protocol=Pretrain + Finetune, Modality=3-stream2023.02 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| LA-GCN2023.05 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MotionBERTTraining protocol=finetune2022.10 | 97.2 | — | — | — | — | — | — | — | — | — | — | 3.1 | — | |
| DSTformerModality=Joint2024.03 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| LA-GCN2026.03 | 97.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DDGCN2023.05 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Info-GCN2023.05 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PoseC3Dmodality=RGB+Pose, #frames=16+322022.11 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Chi et al.Modality=Skeleton2023.12 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DDGCN2021.07 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Info-GCN2026.03 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| HocSLM2026.03 | 97.1 | — | — | — | — | — | — | — | — | — | — | — | — | |
| RNX3D101-CPose modality=true, RGB modality=true, Pose-guided cropping strategy=true2019.12 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | |
| GAPMode=4 ensemble2022.08 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | |
| SkateFormerMod.=J, Param. (M)=2.03, FLOPs (G)=3.62, Architecture paradigm=ANN-Trans.2026.03 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | |
| GAP2026.03 | 97 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Info-GCNMode=4 ensemble2022.08 | 96.9 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCNModality=Skeleton2021.12 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCNSource=Official performance2022.05 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCN2023.05 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCNmodality=Skeleton, #frames=642022.11 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCNMode=4 ensemble2022.08 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Chen et al.Modality=Skeleton2023.12 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCN2021.07 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCNMod.=All, Param. (M)=1.46, FLOPs (G)=7.88, Architecture paradigm=ANN-GCN2026.03 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| CTR-GCN2026.03 | 96.8 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DTNModality=RGB+Depth, Fusion=Addition2021.12 | 96.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Trivedi and SarvadevabhatlaModality=Skeleton2023.12 | 96.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PSTNet++2026.06 | 96.7 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN+ADGModality=Skeleton2021.12 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DTNModality=RGB+Depth, Fusion=Multiplication2021.12 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN+ADG2023.05 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MST-GCN2023.05 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN+ADGmodality=Skeleton, #frames=3002022.11 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN+ADGMode=4 ensemble2022.08 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MST-GCNMode=4 ensemble2022.08 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN+ADG2021.07 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-GCN + ADG2026.03 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MST-GCN2026.03 | 96.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCNModality=Skeleton2021.12 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UNFGEF2020.03 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCNEvaluation protocol=Pretrain + Finetune2023.02 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCN2023.05 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCNmodality=Skeleton, #frames=3002022.11 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMDRmodality=Depth, #frames=642022.11 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| 4S-Shift-GCNMode=4 ensemble2022.08 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PoseC3DModality=Joint2024.03 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCN2021.07 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCNMod.=All, FLOPs (G)=10.00, Architecture paradigm=ANN-GCN2026.03 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Hyperformer (4 ensemble)Mod.=All, Param. (M)=2.60, FLOPs (G)=59.20, Architecture paradigm=ANN-Trans.2026.03 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Shift-GCN2026.03 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PSTNet2026.06 | 96.5 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Sym-GNNvariant=Complete model2019.10 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| P4TransformerModality=point2021.12 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DSTAEvaluation protocol=Pretrain + Finetune2023.02 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| IIP-TransformerEvaluation protocol=Pretrain + Finetune2023.02 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| P4Transformermodality=Point clouds2022.11 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DSTAMode=4 ensemble2022.08 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| P4Transformer2026.06 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| PST-Transformer2026.06 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Kinet2026.06 | 96.4 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BAGCNStream=Spatial2019.12 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ST-TR2023.05 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UMDRmodality=RGB, #frames=642022.11 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MCCTraining protocol=finetune2022.10 | 96.3 | — | — | — | — | — | — | — | — | — | — | 1.2 | — | |
| ST-TR-agcn2026.03 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| 3DV-PointNet++2026.06 | 96.3 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-AAGCNYear=20202020.07 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-G3D NetYear=20202020.07 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-AAGCNPose modality=true, RGB modality=false, Pose-guided cropping strategy=false2019.12 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-AAGCNmodality=multi-stream2019.12 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DTNModality=Depth2021.12 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| DTNModality=Depth2021.12 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-G3DEvaluation protocol=Pretrain + Finetune2023.02 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-G3D2023.05 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Decouple+Recouplemodality=Depth, #frames=642022.11 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| MS-G3DMode=2 ensemble2022.08 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Liu et al.Modality=Skeleton2023.12 | 96.2 | — | — | — | — | — | — | — | — | — | — | — | — |