Research map / 3D vision / Face and shape
Learning to Reconstruct Shapes from Unseen Classes
Read the paperCheck 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 |
|---|---|---|---|
| Disentangle geometric projections from learned shape reconstruction using spherical-map completion | GenRe · 10 unseen ShapeNet classes (Avg) | Chamfer distance: .115 -> .106 | Helped |
| Use two-step spherical-map prediction and inpainting instead of direct RGB-to-final-spherical-map prediction (ablation) | GenRe · 10 unseen ShapeNet classes (Avg) | Chamfer distance: .115 -> .110 | Helped |
| Provide ground-truth depth instead of estimated depth (ablation) | GenRe · 10 unseen ShapeNet classes (Avg) | Chamfer distance: .106 -> .057 | Helped |
| Provide ground-truth full spherical maps (ablation) | GenRe · 10 unseen ShapeNet classes (Avg) | Chamfer distance: .057 -> .036 | Helped |
| Use GenRe rather than the strongest listed competitor for each unseen Pix3D class | GenRe · Unseen Pix3D classes except Bed | Chamfer distance: .109 -> .101 (Bookcase); .121 -> .109 (Desk); .088 -> .083 (Sofa); .124 -> .116 (Table); .116 -> .109 (Wardrobe) | Helped |
| Use GenRe rather than Shin et al. on beds | GenRe · Bed (Pix3D) | Chamfer distance: .106 -> .113 | Hurt |
Closest papers by meaning
- Fostering Generalization in Single-View 3D Reconstruction by Learning a Hierarchy of Local and Global Shape Priors (CVPR 2021 · similarity 0.92)
- Front2Back: Single View 3D Shape Reconstruction via Front to Back Prediction (CVPR 2020 · similarity 0.91)
- PatchComplete: Learning Multi-Resolution Patch Priors for 3D Shape Completion on Unseen Categories (NeurIPS 2022 · similarity 0.90)
- Learning View Priors for Single-View 3D Reconstruction (CVPR 2019 · similarity 0.90)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.88)
- Multistable Shape from Shading Emerges from Patch Diffusion (NeurIPS 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.