Action Recognition on NTU RGB+D 120 (Cross-Subject)
94.85AccuracyENConvLSTM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ENConvLSTM2025.01 | 94.85 | — | |
| Inflated ResNet with Hierarchical ClassificationYear=2020, Pose Input=true, RGB Input=true2020.07 | 93.69 | — | |
| VPN++Stream=VPN++ + 3D Poses2021.05 | 92.5 | — | |
| EPP-NetModality=Skeleton+Parsing2024.01 | 91.1 | — | |
| EPP-Net (J+B+P)Modality=Skeleton+Parsing2024.01 | 90.9 | — | |
| LA-GCNModality=Skeleton+Text, Source=arXiv'232024.01 | 90.7 | — | |
| FreqMixFormerFramework=Transformer, Stream Configuration=6-stream2024.07 | 90.5 | — | |
| DVANetModality=RGB2023.12 | 90.4 | — | |
| STAR-TransformerModality=Skeleton+RGB, Source=WACV'232024.01 | 90.3 | — | |
| STAR-TransformerFramework=Transformer2024.07 | 90.3 | — | |
| BlockGCNPublisher=CVPR20242025.01 | 90.3 | — | |
| FreqMixFormerFramework=Transformer, Stream Configuration=4-stream2024.07 | 90.2 | — | |
| HD-GCNFramework=GCN, Stream Configuration=6-stream2024.07 | 90.1 | — | |
| SkeMixFormerFramework=Transformer, Stream Configuration=6-stream2024.07 | 90.1 | — | |
| ST-GCNRef=[23]2025.01 | 90.1 | — | |
| KoopmanFramework=GCN2024.07 | 90 | — | |
| STEP-CATFormerFramework=Transformer2024.07 | 90 | — | |
| SkeMixFormerFramework=Transformer, Stream Configuration=4-stream2024.07 | 90 | — | |
| LSTModality=Skeleton+Text, Source=ICCV'232024.01 | 89.9 | — | |
| HyperformerFramework=Transformer2024.07 | 89.9 | — | |
| Hyperformer (4 ensemble)Mod.=All, Param. (M)=2.60, FLOPs (G)=59.20, Architecture paradigm=ANN-Trans.2026.03 | 89.9 | — | |
| Chi et al. (1)Modality=Skeleton2023.12 | 89.8 | — | |
| InfoGCNModality=Skeleton, Source=CVPR 222024.01 | 89.8 | — | |
| InfoGCNFramework=GCN, Stream Configuration=6-stream2024.07 | 89.8 | — | |
| HD-GCNFramework=GCN, Stream Configuration=4-stream2024.07 | 89.8 | — | |
| SkateFormer (4 ensemble)Mod.=All, Param. (M)=2.03, FLOPs (G)=14.48, Architecture paradigm=ANN-Trans.2026.03 | 89.8 | — | |
| Stream-GCNFramework=GCN2024.07 | 89.7 | — | |
| DG-STGCNPose Source=3D Skeletons, Fusion=Ensemble (4 modalities)2022.10 | 89.6 | — | |
| FR-HeadFramework=GCN2024.07 | 89.5 | — | |
| Trivedi and SarvadevabhatlaModality=Skeleton2023.12 | 89.4 | — | |
| PSUMNetModality=Skeleton, Source=ECCV'222024.01 | 89.4 | — | |
| InfoGCNFramework=GCN, Stream Configuration=4-stream2024.07 | 89.4 | — | |
| TranSkeletonFramework=Transformer2024.07 | 89.4 | — | |
| NAS-GCN2025.01 | 89.4 | — | |
| LG-SGNetPublisher=PR20252025.01 | 89.4 | — | |
| VPN++Stream=Distilled2021.05 | 89.3 | — | |
| 2s DG-STGCNPose Source=3D Skeletons, Fusion=2-stream (joint+bone)2022.10 | 89.2 | — | |
| 2SAGC-LSTM2025.01 | 89.2 | — | |
| ST-GCNPublisher=AAAI20182025.01 | 89.2 | — | |
| DSTSA-GCNModality=Fusion2025.01 | 89.12 | — | |
| PSTNet++ MotionPDE2026.06 | 89.11 | — | |
| PST-Transformer MotionPDE2026.06 | 89.11 | — | |
| FG-STFormerFramework=Transformer2024.07 | 89 | — | |
| SIT-MLPFramework=Transformer2024.07 | 89 | — | |
| CTR-GCNPose Source=3D Skeletons2022.10 | 88.9 | — | |
| Chen et al.Modality=Skeleton2023.12 | 88.9 | — | |
| CTR-GCNModality=Skeleton, Source=ICCV'212024.01 | 88.9 | — | |
| CTR-GCNFramework=GCN2024.07 | 88.9 | — | |
| CTR-GCNPublisher=ICCV20212025.01 | 88.9 | — | |
| CTR-GCNMod.=All, Param. (M)=1.46, FLOPs (G)=7.88, Architecture paradigm=ANN-GCN2026.03 | 88.9 | — | |
| PSTNet MotionPDE2026.06 | 88.72 | — | |
| Song et al.Modality=Skeleton2023.12 | 88.7 | — | |
| SAN-GCNPublisher=TMM20232025.01 | 88.7 | — | |
| ST-GCN++Pose Source=3D Skeletons2022.10 | 88.6 | — | |
| Duan et al.Modality=Skeleton2023.12 | 88.6 | — | |
| PSTNet++Reference=IEEE TPAMI 222026.06 | 88.6 | — | |
| Chi et al. (2)Modality=Skeleton2023.12 | 88.5 | — | |
| STSA-NetModality=Skeleton, Source=Neuro.'232024.01 | 88.5 | — | |
| 2S-AGCN2025.01 | 88.5 | — | |
| EfficientGCN-B4Framework=GCN2024.07 | 88.3 | — | |
| VPN-AStream=Distilled2021.05 | 88 | — | |
| PointATAInput=Point2026.02 | 88 | — | |
| FreqMixFormerFramework=Transformer, Stream Configuration=joint-only2024.07 | 87.9 | — | |
| VPNStream=Attention Fusion2021.05 | 87.8 | — | |
| SRENet2026.06 | 87.8 | — | |
| SkateFormerMod.=J, Param. (M)=2.03, FLOPs (G)=3.62, Architecture paradigm=ANN-Trans.2026.03 | 87.7 | — | |
| VPN-FStream=Distilled2021.05 | 87.6 | — | |
| DG-STGCNPose Source=2D pose estimations, Fusion=Ensemble (4 modalities)2022.10 | 87.5 | — | |
| MST-GCNModality=Skeleton, Source=AAAI'212024.01 | 87.5 | — | |
| PST-Trans.Input=Point2026.02 | 87.5 | — | |
| PST-TransformerReference=IEEE TPAMI 232026.06 | 87.5 | — | |
| PST-Transformer2026.06 | 87.5 | — | |
| 2s DG-STGCNPose Source=2D pose estimations, Fusion=2-stream (joint+bone)2022.10 | 87.3 | — | |
| DynamicGCNModality=Skeleton, Source=MM'202024.01 | 87.3 | — | |
| Dynamic GCNPublisher=MM20202025.01 | 87.3 | — | |
| MS-G3D++Pose Source=2D pose estimations2022.10 | 87.2 | — | |
| SkeMixFormerFramework=Transformer, Stream Configuration=joint-only2024.07 | 87.1 | — | |
| PSTNetInput=point2022.05 | 87 | — | |
| TSMFModality=Skeleton+RGB, Source=AAAI'212024.01 | 87 | — | |
| PSTNetInput=Point2026.02 | 87 | — | |
| PSTNetReference=ICLR 212026.06 | 87 | — | |
| PSTNet2026.06 | 87 | — | |
| MS-G3D NetYear=2020, Pose Input=true, RGB Input=false2020.07 | 86.9 | — | |
| MS-G3DPose Source=3D Skeletons2022.10 | 86.9 | — | |
| PoseC3DPose Source=2D pose estimations2022.10 | 86.9 | — | |
| Liu et al.Modality=Skeleton2023.12 | 86.9 | — | |
| MS-G3DModality=Skeleton, Source=CVPR'202024.01 | 86.9 | — | |
| MS-G3DFramework=GCN2024.07 | 86.9 | — | |
| STGR-GCN2025.01 | 86.9 | — | |
| MS-G3DMod.=J+B, Param. (M)=3.19, FLOPs (G)=48.88, Architecture paradigm=ANN-GCN2026.03 | 86.9 | — | |
| AS-GCN2025.01 | 86.8 | — | |
| DSTSA-GCNModality=Bone2025.01 | 86.8 | — | |
| DSTA-NetPose Source=3D Skeletons2022.10 | 86.6 | — | |
| DSTA-NetModality=Skeleton, Source=ACCV'202024.01 | 86.6 | — | |
| DSTAFramework=Transformer2024.07 | 86.6 | — | |
| HyperformerMod.=J, Param. (M)=2.60, FLOPs (G)=14.80, Architecture paradigm=ANN-Trans.2026.03 | 86.6 | — | |
| ST-GCN++Pose Source=2D pose estimations2022.10 | 86.4 | — | |
| 4s-GSTNFramework=Transformer2024.07 | 86.4 | — | |
| P4Trans.Input=Point2026.02 | 86.4 | — | |
| P4TransformerReference=CVPR 212026.06 | 86.4 | — |