Group Activity Recognition 벤치마크
Group Activity Recognition on Collective Activity
Accuracy
- 2015-11-13 — Deng et al.: Accuracy 81.2
- 2018-11-01 — H-LSTCM: Accuracy 83.75
- 2019-04-23 — GT (Inception-v3): Accuracy 91.0
- 2021-12-11 — COMPOSER: Accuracy 96.2
- 2022-07-12 — Tamura et al.: Accuracy 96.5
| Rank | Model | Accuracy | Extra Training Data | Paper | Code | Year |
|---|---|---|---|---|---|---|
| 1 | Tamura et al. | 96.5 | ✓ | Hunting Group Clues with Transformers for Social Group Activity Recognition | 2022 | |
| 2 | COMPOSER | 96.2 | COMPOSER: Compositional Reasoning of Group Activity in Videos with Keypoint-Only Modality | hongluzhou/composer | 2021 | |
| 3 | GT (Inception-v3) | 91 | Learning Actor Relation Graphs for Group Activity Recognition | wjchaoGit/Group-Activity-Recognition · East-Tree/groupAc_GCN | 2019 | |
| 4 | CHASE(CTR-GCN) | 89.61 | CHASE: Learning Convex Hull Adaptive Shift for Skeleton-based Multi-Entity Action Recognition | Necolizer/CHASE | 2024 | |
| 5 | H-LSTCM | 83.75 | Hierarchical Long Short-Term Concurrent Memory for Human Interaction Recognition | 2018 | ||
| 6 | Deng et al. | 81.2 | Structure Inference Machines: Recurrent Neural Networks for Analyzing Relations in Group Activity Recognition | 2015 |