Research map / 3D vision / Face and shape
Structured Prediction of Unobserved Voxels From a Single Depth Image
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 |
|---|---|---|---|
| Use Voxlets with observed-fit selection rather than the ground-truth-segmented Zheng baseline | Tabletop volumetric completion · Tabletop test scenes | IoU: 0.528 -> 0.585 | Helped |
| Use Voxlets rather than ground-truth-segmented bounding boxes | Volumetric completion · Synthetic NYU-Depth V2 test scenes | IoU: 0.349 -> 0.508 | Helped |
| Select the proposal most consistent with observed geometry rather than the forest medoid (ablation) | Voxlets · Tabletop test scenes | IoU: 0.326 -> 0.585 | Helped |
| Select the proposal most consistent with observed geometry rather than averaging forest predictions (ablation) | Voxlets · Tabletop test scenes | IoU: 0.312 -> 0.585 | Helped |
| Add observed-fit selection rather than forest medoid selection (ablation) | Voxlets · Tabletop test scenes | Recall: 0.358 -> 0.658 | Helped |
| Add observed-fit selection rather than forest mean selection (ablation) | Voxlets · Tabletop test scenes | Recall: 0.337 -> 0.658 | Helped |
| Compress and decompress the local ground-truth voxels with PCA (ablation) | Voxlets support-region representation · Tabletop test scenes | IoU: 0.962 -> 0.908 | Hurt |
Closest papers by meaning
- Pixels, Voxels, and Views: A Study of Shape Representations for Single View 3D Object Shape Prediction (CVPR 2018 · similarity 0.91)
- Semantic Multi-View Stereo: Jointly Estimating Objects and Voxels (CVPR 2017 · similarity 0.89)
- Learning 3D Object Shape and Layout Without 3D Supervision (CVPR 2022 · similarity 0.89)
- Geometry-Aware Network for Non-Rigid Shape Prediction From a Single View (CVPR 2018 · similarity 0.89)
- Large Spatial Model: End-to-end Unposed Images to Semantic 3D (NeurIPS 2024 · similarity 0.87)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.87)
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.