| Paper | Code | Average accuracy of 3 splits | Accuracy | No. parameters | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|
| Hierarchical Temporal Convolution Network:Towards Privacy-Centric Activity Recognition | ✓ Link | 86.1 | 86.1 | 1.75 | HT-ConvNet | 2024-12-21 |
| Make Skeleton-based Action Recognition Model Smaller, Faster and Better | ✓ Link | 77.2 | 78.0 (average of 3 split train/test) | 1.82 M | DD-Net | 2019-07-23 |
| Pose And Joint-Aware Action Recognition | ✓ Link | 68.55 | JMRN (No GT pose) | 2020-10-16 | ||
| PoTion: Pose MoTion Representation for Action Recognition | 67.9 | - | PoTion | 2018-06-01 | ||
| Simple yet efficient real-time pose-based action recognition | ✓ Link | 65.5 | EHPI | 2019-04-19 | ||
| Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection | ✓ Link | 56.8 | Chained | 2017-04-03 |