Research map / 3D vision / Face and shape
Learning Shape Priors for Single-View 3D Completion and Reconstruction
Jiajun Wu, Chengkai Zhang, Xiuming Zhang, Zhoutong Zhang, William T. Freeman, Joshua B. Tenenbaum
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 naturalness loss to ShapeHD completion (ablation) | ShapeHD · ShapeNet completion 32^3 avg | Chamfer distance: .093 -> .084 | Improved |
| add naturalness loss (ablation) | ShapeHD · ShapeNet completion 32^3 avg | IoU: .529 -> .529 | No clear effect |
| add naturalness loss (ablation) | ShapeHD · ShapeNet completion 32^3 chair | IoU: .466 -> .488 | Improved |
| add naturalness loss (ablation) | ShapeHD · ShapeNet completion 32^3 plane | IoU: .488 -> .452 | Hurt |
| add naturalness loss (ablation) | ShapeHD · ShapeNet completion 32^3 chair | Chamfer distance: .112 -> .096 | Improved |
| ShapeHD instead of 3D-EPN | ShapeHD · ShapeNet completion 32^3 avg | IoU: .181 -> .529 | Improved |
| ShapeHD instead of DRC (3D) | ShapeHD · Pix3D reconstruction | Chamfer distance: 0.160 -> 0.123 | Improved |
| ShapeHD instead of OGN on cars | ShapeHD · PASCAL 3D+ reconstruction car | Chamfer distance: 0.087 -> 0.129 | Hurt |
| ShapeHD instead of DRC (3D) | ShapeHD · PASCAL 3D+ reconstruction avg | Chamfer distance: 0.122 -> 0.119 | Improved |
| joint training of naturalness module from scratch instead of pretrain then fine-tune (ablation) | ShapeHD · training stability | stability | Hurt |
Closest papers by meaning
- Multimodal Shape Completion via Conditional Generative Adversarial Networks (ECCV 2020 · similarity 0.91)
- Fostering Generalization in Single-View 3D Reconstruction by Learning a Hierarchy of Local and Global Shape Priors (CVPR 2021 · similarity 0.91)
- PatchComplete: Learning Multi-Resolution Patch Priors for 3D Shape Completion on Unseen Categories (NeurIPS 2022 · similarity 0.91)
- Learning to Reconstruct Shapes from Unseen Classes (NeurIPS 2018 · similarity 0.90)
- Diverse Shape Completion via Style Modulated Generative Adversarial Networks (NeurIPS 2023 · similarity 0.90)
- 3D Shape Completion with Multi-View Consistent Inference (AAAI 2020 · similarity 0.90)
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.
