Research map / Optimization / Optimizers
Fits Like a Glove: Rapid and Reliable Hand Shape Personalization
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| track with personalized hand shape model instead of template (temporal tracking, S4 vs S3) (ablation) | Hand tracker of Sharp et al. · NYU Hand Pose | marker localization error (% of frames under threshold) | Helped |
| personalize model but disable temporal coherence (S2 vs S1, per-frame) (ablation) | Hand tracker of Sharp et al. · NYU Hand Pose | marker localization error | Mixed |
| personalized vs template model tracking on a single-subject sequence set (ablation) | Hand tracker of Sharp et al. · Dexter | fingertip marker localization error | Mixed |
| increase shape basis dimension K from 1 to 5 for personalization (ablation) | personalized tracker · FingerPaint | pixel classification accuracy | Helped |
| personalize with K=1 (scaled mean hand) versus hand-made template (ablation) | personalized tracker · FingerPaint | pixel classification accuracy | No clear effect |
| add pose prior to template tracking of Sharp et al. (ablation) | template tracker · FingerPaint | pixel classification accuracy | Helped |
| remove the pose prior from the lifted energy | shape fitting energy · self-occluded poses | rank of J and J^T J | Hurt |
| use larger rendered tile size (64x64 vs 512x512) for the golden energy (ablation) | golden energy · x-axis translation sweep | scale of discontinuities | Helped |
| optimize shape and poses with LM and central-difference gradients from the mean shape | shape fitting · synthetic data, F=40 | energy E' and |beta1 - beta1_gt| | Helped |
Closest papers by meaning
- LISA: Learning Implicit Shape and Appearance of Hands (CVPR 2022 · similarity 0.86)
- Efficiently Creating 3D Training Data for Fine Hand Pose Estimation (CVPR 2016 · similarity 0.86)
- URHand: Universal Relightable Hands (CVPR 2024 · similarity 0.84)
- FoundHand: Large-Scale Domain-Specific Learning for Controllable Hand Image Generation (CVPR 2025 · similarity 0.84)
- SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-training (ICLR 2025 · similarity 0.84)
- SRHand: Super-Resolving Hand Images and 3D Shapes via View/Pose-aware Neural Image Representations and Explicit Meshes (NeurIPS 2025 · 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.