Research map / 3D vision / Human pose estimation
FLAG3D: A 3D Fitness Activity Dataset With Language Instruction
Yansong Tang, Jinpeng Liu, Aoyang Liu, Bin Yang, Wenxun Dai, Yongming Rao, Jiwen Lu, Jie Zhou, Xiu Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Transfer skeleton action recognition from rendered videos to natural videos | 2s-AGCN · FLAG3D | accuracy: 98.6 -> 81.6 | Hurt |
| Pretrain on FLAG3D before transfer learning | ST-GCN · FineGym | accuracy: 91.4 / 92.0 (+0.6) | Improved |
| Pretrain on FLAG3D before transfer learning | ST-GCN · NTU60-XSub | accuracy: 89.0 / 89.0 (+0.0) | No clear effect |
| Pretrain on FLAG3D before transfer learning | 2s-AGCN · FineGym | accuracy: 91.8 / 92.1 (+0.3) | Improved |
| Pretrain on FLAG3D before transfer learning | 2s-AGCN · NTU60-XSub | accuracy: 89.7 / 91.0 (+1.3) | Improved |
| Pretrain on FLAG3D before transfer learning | MS-G3D · FineGym | accuracy: 92.7 / 93.4 (+0.7) | Improved |
| Pretrain on FLAG3D before transfer learning | MS-G3D · NTU60-XSub | accuracy: 92.2 / 92.3 (+0.1) | Improved |
| Fine-tune on FLAG3D | ROMP · FLAG3D | MPJPE: 379.44 -> 114.73 | Improved |
| Fine-tune on FLAG3D | ROMP · FLAG3D | PA-MPJPE: 100.48 -> 62.29 | Improved |
| Fine-tune on FLAG3D | ROMP · Challenging cases P2 | MPJPE: 490.248 -> 131.001 | Improved |
| Pretrain on FLAG3D | MDM · KIT | R-Precision: 0.396 -> 0.407 | Improved |
| Pretrain on FLAG3D | MDM · KIT | FID: 0.497 -> 0.491 | Improved |
Closest papers by meaning
- FLAG3D++: A Benchmark for 3D Fitness Activity Comprehension With Language Instruction (TPAMI 2025 · similarity 0.95)
- M3GYM: A Large-Scale Multimodal Multi-view Multi-person Pose Dataset for Fitness Activity Understanding in Real-world Settings (CVPR 2025 · similarity 0.85)
- Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives (CVPR 2024 · similarity 0.84)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.84)
- Learning from Synthetic Human Group Activities (CVPR 2024 · similarity 0.84)
- MVHumanNet: A Large-scale Dataset of Multi-view Daily Dressing Human Captures (CVPR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
