Research map / 3D vision / Human pose estimation
Single View Physical Distance Estimation Using Human Pose
Xiaohan Fei, Henry Wang, Lin Lee Cheong, Xiangyu Zeng, Meng Wang, Joseph Tighe
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 direct linear solver instead of line intersection and fitting | pose-based calibration · Simulation noise std 0.5 px | focal length error f_x (%): 4.58 -> 3.11 | Improved |
| use direct linear solver instead of line intersection and fitting | pose-based calibration · Simulation noise std 5.0 px | focal length error f_x (%): 32.69 -> 28.03 | Improved |
| use direct linear solver instead of line intersection and fitting | pose-based calibration · Simulation noise std 5.0 px | failure rate (%): 27.28 -> 13.72 | Improved |
| aggregate more people measurements | direct linear solver · Simulation, number of people | failure rate (%): 9.88 -> 1.52 | Improved |
| aggregate more people measurements | direct linear solver · Simulation, number of people | 3-D point error X (%): 16.31 -> 9.31 | Improved |
| add lens distortion modeling (ablation) | direct linear solver · Simulation k1=1e-3 | focal length error f_x (%): 0.16 -> 1.077e-5 | Improved |
| add lens distortion modeling (ablation) | direct linear solver · Simulation k1=-1e-3 | focal length error f_y (%): 0.020 -> 0.0019 | Improved |
| add lens distortion modeling (ablation) | direct linear solver · Simulation noisy measurements | failure rate | Hurt |
| use pose-based geometric estimator instead of fine-tuned ResNet classifier | distance estimation · MEVADA test set | accuracy: 0.40 -> 0.78 | Improved |
| use direct solver instead of prior line-based method | calibration · vPTZ set1-cam-132 | focal length error (%): 10.14% -> 0.35 % | Improved |
| use direct solver instead of prior line-based method | calibration · POM terrace1-cam0 | focal length error (%): 1.43% -> 2.51% | Hurt |
Closest papers by meaning
- Camera Distance-Aware Top-Down Approach for 3D Multi-Person Pose Estimation From a Single RGB Image (ICCV 2019 · similarity 0.86)
- A Comparative Assessment of Accuracy in Video-Based Monocular Human Pose Estimation Frameworks (TPAMI 2026 · similarity 0.86)
- Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes - The Importance of Multiple Scene Constraints (CVPR 2018 · similarity 0.86)
- FreeMan: Towards Benchmarking 3D Human Pose Estimation under Real-World Conditions (CVPR 2024 · similarity 0.85)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.85)
- MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation (CVPR 2025 · similarity 0.85)
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.
