Research map / 3D vision / Human pose estimation
PACE: Pose Annotations in Cluttered Environments
Yang You, kai xiong, Zhening Yang, Zhengxiang Huang, Junwei Zhou, Ruoxi Shi, Zhou FANG, Adam Harley, Leonidas Guibas, Cewu Lu
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 |
|---|---|---|---|
| evaluate SOTA category-level methods on PACE instead of NOCS-REAL275 | HS-Pose, DualPoseNet, SARNet · bottle, bowl, can, mug | AP@15deg5cm: 99.8 -> 0.4 (HS-Pose bottle) | Hurt |
| use classical PPF instead of deep learning methods for instance-level pose estimation | PACE instance-level benchmark · PACE | AR: 9.0 -> 52.2 | Improved |
| train on simulated data instead of real data (single instance/category) (ablation) | GDRNPP · PACE instance-level | AR (%): 88.2 -> 78.1 | Hurt |
| train on simulated data instead of real data (single category) (ablation) | HS-Pose · PACE category-level | mean AP@0:60deg,0:15cm: 81.2 -> 1.1 | Hurt |
| scale training from a single instance to all instances (ablation) | GDRNPP · PACE instance-level | AR (%): 88.2 -> 9.7 | Hurt |
| scale training from a single category to all categories (ablation) | CPPF++ · PACE category-level | mean AP@0:60deg,0:15cm: 83.4 -> 13.7 | Hurt |
| apply rigid-object-focused category-level methods to articulated objects | category-level methods on PACE · PACE articulated objects | AP | Hurt |
| use depth-based model ICG instead of RGB-only RBOT for model-based tracking | model-based tracking · PACE rigid | ADD(-S) AUC: 7.4 -> 38.1 | Improved |
Closest papers by meaning
- TEMPO: Efficient Multi-View Pose Estimation, Tracking, and Forecasting (ICCV 2023 · similarity 0.87)
- Slim DensePose: Thrifty Learning From Sparse Annotations and Motion Cues (CVPR 2019 · similarity 0.86)
- Beyond Standard Benchmarks: Parameterizing Performance Evaluation in Visual Object Tracking (ICCV 2017 · similarity 0.86)
- MotionMap: Representing Multimodality in Human Pose Forecasting (CVPR 2025 · similarity 0.85)
- Dynamic Camera Poses and Where to Find Them (CVPR 2025 · similarity 0.85)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · 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.
