Research map / 3D vision / Human pose estimation
Gaussian-Mixture Latent Flow for Stochastic 3D Human Motion Prediction
Yue Ma, Frederick W. B. Li, Xiaohui Liang
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 |
|---|---|---|---|
| learn the mixture prior instead of fixed sub-Gaussians (ablation) | Gaussian-mixture latent flow · Human3.6M | FID: 0.133 -> 0.088 | Improved |
| learn the mixture prior instead of fixed sub-Gaussians (ablation) | Gaussian-mixture latent flow · Human3.6M | CMD: 3.015 vs 2.767 | Hurt |
| use Gaussian-mixture prior instead of standard normal (ablation) | Gaussian-mixture latent flow · Human3.6M | CMD: 5.854 -> 3.015 | Improved |
| use Gaussian-mixture prior instead of standard normal (ablation) | Gaussian-mixture latent flow · Human3.6M | ADE: 0.356 -> 0.333 | Improved |
| add the latent space with mixture prior versus flow matching in data space (ablation) | Gaussian-mixture latent flow · AMASS | CMD: 12.563 -> 8.579 | Improved |
| use past-informed source distribution N(Z0,I) instead of N(0,I) (ablation) | Gaussian-mixture latent flow · Human3.6M | FID: 0.223 -> 0.088 | Improved |
| condition on observed motion (ablation) | Gaussian-mixture latent flow · Human3.6M | ADE: 0.352 -> 0.333 | Improved |
| use temporal tokenization instead of spatial tokenization (ablation) | Gaussian-mixture latent flow · Human3.6M | ADE: 0.397 -> 0.333 | Improved |
| increase initial mixture components from 8 to 16 (ablation) | Gaussian-mixture latent flow · Human3.6M | FID: 0.134 -> 0.088 | Improved |
| increase initial mixture components from 16 to 32 (ablation) | Gaussian-mixture latent flow · Human3.6M | FID: 0.088 -> 0.092 | No clear effect |
| use learnable single-component prior instead of standard normal (ablation) | Gaussian-mixture latent flow · Human3.6M | CMD: 5.854 -> 3.215 | Improved |
| increase Euler steps from 10 to 100 (ablation) | Gaussian-mixture latent flow · Human3.6M | APD: 3.427 -> 4.804 | Improved |
Closest papers by meaning
- BeLFusion: Latent Diffusion for Behavior-Driven Human Motion Prediction (ICCV 2023 · similarity 0.89)
- FlowMoP: Stochastic Multi-Person Motion Prediction (NeurIPS 2026 · similarity 0.89)
- Harmonizing Stochasticity and Determinism: Scene-responsive Diverse Human Motion Prediction (NeurIPS 2024 · similarity 0.88)
- Nonisotropic Gaussian Diffusion for Realistic 3D Human Motion Prediction (CVPR 2025 · similarity 0.88)
- Existence Is Chaos: Enhancing 3D Human Motion Prediction with Uncertainty Consideration (AAAI 2024 · similarity 0.87)
- Motion Diversification Networks (CVPR 2024 · similarity 0.87)
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.
