Research map / 3D vision / Human pose estimation
Tracking and Reconstructing Hand Object Interactions from Point Cloud Sequences in the Wild
Jiayi Chen, Mi Yan, Jiazhao Zhang, Yinzhen Xu, Xiaolong Li, Yijia Weng, Li Yi, Shuran Song, He Wang
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 hand-object optimization (ablation) | HandTrackNet tracking and MANO reconstruction · HO3D | MPJPE (cm): 2.34 -> 2.06 | Improved |
| add hand-object optimization (ablation) | HandTrackNet tracking and MANO reconstruction · DexYCB | MPJPE (cm): 2.99 -> 2.69 | Improved |
| train with simulated sensor depth rather than perfect depth (ablation) | HandTrackNet · HO3D, DexYCB, and SimGrasp | tracking performance | Improved |
| use joint-position-based canonicalization instead of no canonicalization or OBB (ablation) | HandTrackNet · real-world hand tracking | tracking performance | Improved |
| use previous-frame joints for the coarse joint estimate (ablation) | HandTrackNet · hand tracking under occlusion | tracking performance | Improved |
| communicate joint features across neighboring fingers (ablation) | HandTrackNet rearrangement layer · hand tracking | tracking performance | Improved |
| update hand shape during tracking (ablation) | Ours · HO3D | MPJPE (cm): 2.34 -> 2.26 | Improved |
| update hand shape during tracking (ablation) | Ours · DexYCB | MPJPE (cm): 2.99 -> 2.89 | Improved |
| update object shape from accumulated point clouds (ablation) | Ours object tracker · HO3D bottle | CD (cm): 2.06 -> 1.79 | Improved |
| update object shape from accumulated point clouds (ablation) | Ours object tracker · DexYCB bowl | CD (cm): 4.71 -> 5.31 | Hurt |
| use three KNN query and rearrangement repetitions instead of two (ablation) | HandTrackNet · HO3D | MPJPE (cm): 2.11 -> 2.03 | Improved |
| use three KNN query and rearrangement repetitions instead of two (ablation) | HandTrackNet · DexYCB | MPJPE (cm): 2.75 -> 2.62 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- CAPTRA — CAPTRA: CAtegory-Level Pose Tracking for Rigid and Articulated Objects From Point Clouds (ICCV 2021) · better on 2 of 2 shared results
- HandFoldingNet — HandFoldingNet: A 3D Hand Pose Estimation Network Using Multiscale-Feature Guided Folding of a 2D Hand Skeleton (ICCV 2021) · 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
- ComPose: When to Trust Hands for Object Pose Tracking (NeurIPS 2026 · similarity 0.90)
- Real-Time Joint Tracking of a Hand Manipulating an Object from RGB-D Input (ECCV 2016 · similarity 0.89)
- BigHand2.2M Benchmark: Hand Pose Dataset and State of the Art Analysis (CVPR 2017 · similarity 0.89)
- EasyHOI: Unleashing the Power of Large Models for Reconstructing Hand-Object Interactions in the Wild (CVPR 2025 · similarity 0.89)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.89)
- HOISDF: Constraining 3D Hand-Object Pose Estimation with Global Signed Distance Fields (CVPR 2024 · similarity 0.88)
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.
