Research map / 3D vision / Human pose estimation
Adversarial Geometry-Aware Human Motion Prediction
Liang-Yan Gui, Yu-Xiong Wang, Xiaodan Liang, Jose M. F. Moura
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 |
|---|---|---|---|
| use geodesic loss instead of Euclidean loss (ablation) | residual sup. predictor · H3.6M Walking 400ms | mean angle error: 0.81 -> 0.73 | Improved |
| use geodesic loss instead of Euclidean loss (ablation) | residual sup. predictor · H3.6M average 400ms | mean angle error: 1.15 -> 1.07 | Improved |
| add both fidelity and continuity discriminators to geodesic loss (ablation) | AGED · H3.6M Walking 400ms | mean angle error: 0.73 -> 0.67 | Improved |
| add both discriminators to geodesic loss (ablation) | AGED · H3.6M Smoking 400ms | mean angle error: 0.99 -> 0.84 | Improved |
| add both discriminators to geodesic loss (ablation) | AGED · H3.6M Discussion 400ms | mean angle error: 1.06 -> 0.83 | Improved |
| add both discriminators to geodesic loss (ablation) | AGED · H3.6M average 400ms | mean angle error: 1.07 -> 0.97 | Improved |
| use geodesic loss instead of Euclidean loss with both discriminators (ablation) | AGED · H3.6M average 400ms | mean angle error: 1.00 -> 0.97 | Improved |
| use geodesic loss instead of Euclidean loss with both discriminators (ablation) | AGED · H3.6M Discussion 400ms | mean angle error: 0.87 -> 0.83 | Improved |
| add discriminators and geodesic loss over residual sup. | AGED · H3.6M Smoking 80ms vs zero-velocity | mean angle error: 0.26 -> 0.27 | Improved |
| add geodesic loss and discriminators to residual sup. | AGED · H3.6M long-term Walking 1000ms | mean angle error: 1.03 -> 0.91 | Improved |
| add geodesic loss and discriminators to residual sup. | AGED · H3.6M long-term Discussion 1000ms | mean angle error: 1.69 -> 1.30 | Improved |
| use quaternion distance instead of the rotation-matrix geodesic distance (ablation) | AGED · H3.6M | mean angle error | Hurt |
Closest papers by meaning
- Structured Prediction Helps 3D Human Motion Modelling (ICCV 2019 · similarity 0.89)
- Learning Dynamic Relationships for 3D Human Motion Prediction (CVPR 2020 · similarity 0.88)
- Towards Natural and Accurate Future Motion Prediction of Humans and Animals (CVPR 2019 · similarity 0.88)
- Investigating Pose Representations and Motion Contexts Modeling for 3D Motion Prediction (TPAMI 2022 · similarity 0.88)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.87)
- Existence Is Chaos: Enhancing 3D Human Motion Prediction with Uncertainty Consideration (AAAI 2024 · similarity 0.86)
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.
