Research map / 3D vision / Human pose estimation
Ultra Diffusion Poser: Diffusion-Based Human Motion Tracking from Sparse Inertial Sensors and Ranging-based Between-sensor Distances
Dominik Hollidt, Tommaso Bendinelli, Christian Holz
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 |
|---|---|---|---|
| add Spatial Layout Module (ablation) | UDP · TotalCapture | JPE (cm): 4.03 -> 3.76 | Improved |
| add UWB-Diffusion Guidance (ablation) | UDP · TotalCapture | JPE (cm): 4.05 -> 3.76 | Improved |
| add both SPL and UWB-Diffusion Guidance (ablation) | UDP · TotalCapture | JPE (cm): 4.39 -> 3.76 | Improved |
| add Rotation Estimator to orient layouts (ablation) | UDP · TotalCapture | JPE (cm): 3.92 -> 3.76 | Improved |
| add residual P_res refinement of layout (ablation) | UDP · TotalCapture | JPE (cm): 3.79 -> 3.76 | No clear effect |
| use sqrt noise schedule instead of cosine (ablation) | UDP · TotalCapture | JPE (cm): 3.87 -> 3.76 | Improved |
| use LSTM denoiser instead of transformer encoder (ablation) | UDP · TotalCapture | JPE (cm): 4.79 -> 3.76 | Improved |
| add forward kinematics loss (ablation) | UDP · TotalCapture | JPE (cm): 3.76 -> 3.89 | Hurt |
| absolute translation instead of velocity representation (ablation) | UDP · TotalCapture | JPE (cm): 3.95 -> 3.76 | Improved |
| provide ground-truth body shape as input (ablation) | UDP · DanceDB | JPE (cm): 4.67 -> 4.30 | Improved |
| allow lookahead (not last-frame-only), 1 s window (ablation) | UDP · TotalCapture | JPE (cm): 4.14 -> 3.84 | Improved |
| lengthen window from 1 s to 3 s (ablation) | UDP · TotalCapture | JPE (cm): 3.84 -> 3.76 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- UMotion — UMotion: Uncertainty-driven Human Motion Estimation from Inertial and Ultra-wideband Units (CVPR 2025) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiffusionPoser: Real-time Human Motion Reconstruction From Arbitrary Sparse Sensors Using Autoregressive Diffusion (CVPR 2024 · similarity 0.89)
- UMotion: Uncertainty-driven Human Motion Estimation from Inertial and Ultra-wideband Units (CVPR 2025 · similarity 0.89)
- I'M HOI: Inertia-aware Monocular Capture of 3D Human-Object Interactions (CVPR 2024 · similarity 0.85)
- A Unified Diffusion Framework for Scene-aware Human Motion Estimation from Sparse Signals (CVPR 2024 · similarity 0.85)
- Dynamic Inertial Poser (DynaIP): Part-Based Motion Dynamics Learning for Enhanced Human Pose Estimation with Sparse Inertial Sensors (CVPR 2024 · similarity 0.85)
- $\text{Di}^2\text{Pose}$: Discrete Diffusion Model for Occluded 3D Human Pose Estimation (NeurIPS 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.
