Research map / 3D vision / 3D generation
TOCH: Spatio-Temporal Object-to-Hand Correspondence for Motion Refinement
Keyang Zhou, Bharat Lal Bhatnagar, Jan Eric Lenssen, Gerard Pons-Moll
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 |
|---|---|---|---|
| use object-centric TOCH fields instead of hand-centric input (ablation) | TOCH · GRAB-B (0.01 & 0.3) | C-IoU (%): 18.9 -> 26.6 | Improved |
| use object-centric TOCH fields instead of hand-centric input (ablation) | TOCH · GRAB-B (0.01 & 0.3) | MPJPE (mm): 11.2 -> 10.3 | Improved |
| add dense hand-point correspondence to the TOCH field (ablation) | TOCH · GRAB-B (0.01 & 0.3) | C-IoU (%): 18.6 -> 26.6 | Improved |
| add dense hand-point correspondence to the TOCH field (ablation) | TOCH · GRAB-B (0.01 & 0.3) | MPJPE (mm): 12.2 -> 10.3 | Improved |
| add GRU temporal modeling instead of global feature aggregation (ablation) | TOCH · GRAB-B (0.01 & 0.3) | C-IoU (%): 20.4 -> 26.6 | Improved |
| add GRU temporal modeling (ablation) | TOCH · GRAB-B (0.01 & 0.3) | IV (cm^3): 1.87 -> 1.78 | Improved |
| test on objects seen in training (action-based split) instead of unseen objects (ablation) | TOCH · GRAB-B (0.01 & 0.3) | C-IoU (%): 26.6 -> 23.1 | Hurt |
| refine pose-dominant perturbation GRAB-R (0.3) with TOCH | TOCH · GRAB-R (0.3) | MPJPE (mm): 4.58 -> 9.58 | Hurt |
| refine translation-dominant perturbation with TOCH | TOCH · GRAB-T (0.02) | MPJPE (mm): 31.9 -> 12.3 | Improved |
| refine GRAB-R (0.3) with TOCH for contact | TOCH · GRAB-R (0.3) | C-IoU (%): 11.4 -> 26.6 | Improved |
| refine GRAB-T (0.02) with TOCH for penetration | TOCH · GRAB-T (0.02) | IV (cm^3): 2.40 -> 2.50 | No clear effect |
| use TOCH instead of ContactOpt | TOCH · HO-3D | IV (cm^3): 5.71 -> 4.66 | Improved |
Later papers that compared with this method
2 later accepted papers report TOCH as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Hand-Centric Motion Refinement for 3D Hand-Object Interaction via Hierarchical Spatial-Temporal Modeling (AAAI 2024 · better on 7 of 7 shared results)
- ContactGen: Generative Contact Modeling for Grasp Generation (ICCV 2023 · better on 4 of 4 shared results)
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.
- ContactOpt — ContactOpt: Optimizing Contact To Improve Grasps (CVPR 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
- Hand-Centric Motion Refinement for 3D Hand-Object Interaction via Hierarchical Spatial-Temporal Modeling (AAAI 2024 · similarity 0.91)
- ComPose: When to Trust Hands for Object Pose Tracking (NeurIPS 2026 · similarity 0.89)
- How Do I Do That? Synthesizing 3D Hand Motion and Contacts for Everyday Interactions (CVPR 2025 · similarity 0.89)
- HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video (CVPR 2024 · similarity 0.89)
- GEARS: Local Geometry-aware Hand-object Interaction Synthesis (CVPR 2024 · similarity 0.88)
- LatentHOI: On the Generalizable Hand Object Motion Generation with Latent Hand Diffusion. (CVPR 2025 · 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.
