Research map / 3D vision / Face and shape
3D-PRNN: Generating Shape Primitives With Recurrent Neural Networks
Chuhang Zou, Ersin Yumer, Jimei Yang, Duygu Ceylan, Derek Hoiem
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 rotation axis loss to 3D-PRNN (ablation) | 3D-PRNN · ModelNet table (synthetic depth) | shape IoU: 0.188 -> 0.\textbf{263} | Improved |
| add rotation axis loss to 3D-PRNN (ablation) | 3D-PRNN · ModelNet chair (synthetic depth) | shape IoU: 0.245 -> 0.238 | Hurt |
| add rotation axis loss to 3D-PRNN (ablation) | 3D-PRNN · ModelNet night stand (synthetic depth) | shape IoU: 0.204 -> 0.266 | Improved |
| add rotation axis loss to 3D-PRNN (ablation) | 3D-PRNN · ModelNet night stand (synthetic depth) | surface-to-surface distance: 0.104 -> 0.092 | Improved |
| add rotation axis loss to 3D-PRNN (ablation) | 3D-PRNN · ModelNet table (synthetic depth) | surface-to-surface distance: 0.080 -> 0.078 | Improved |
| fine-tune 3D-PRNN on NYU Depth v2 (ablation) | 3D-PRNN · NYU Depth v2 chair (real depth) | shape IoU: 0.138 -> 0.158 | Improved |
| fine-tune 3D-PRNN on NYU Depth v2 (ablation) | 3D-PRNN · NYU Depth v2 table (real depth) | shape IoU: 0.052 -> 0.075 | Improved |
| fine-tune 3D-PRNN on NYU Depth v2 (ablation) | 3D-PRNN · NYU Depth v2 night stand (real depth) | shape IoU: 0.086 -> 0.081 | No clear effect |
| fine-tune 3D-PRNN on NYU Depth v2 (ablation) | 3D-PRNN · NYU Depth v2 chair (real depth) | surface-to-surface distance: 0.110 -> 0.112 | No clear effect |
| fine-tune 3D-PRNN on NYU Depth v2 (ablation) | 3D-PRNN · NYU Depth v2 table (real depth) | surface-to-surface distance: 0.181 -> 0.168 | Improved |
| compare 3D-PRNN with voxel generator of Wu et al. | 3D-PRNN + rot loss · ModelNet night stand (synthetic depth) | surface-to-surface distance: 0.092 vs 0.057 | Hurt |
| compare 3D-PRNN with nearest neighbor retrieval | 3D-PRNN + rot loss · ModelNet table (synthetic depth) | shape IoU: 0.220 -> 0.\textbf{263} | Improved |
Closest papers by meaning
- Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks (CVPR 2021 · similarity 0.90)
- Inferring 3D Shapes from Image Collections Using Adversarial Networks (IJCV 2020 · similarity 0.90)
- SurfNet: Generating 3D Shape Surfaces Using Deep Residual Networks (CVPR 2017 · similarity 0.90)
- Synthesizing 3D Shapes via Modeling Multi-View Depth Maps and Silhouettes With Deep Generative Networks (CVPR 2017 · similarity 0.90)
- Unsupervised Primitive Discovery for Improved 3D Generative Modeling (CVPR 2019 · similarity 0.90)
- From Images to 3D Shape Attributes (TPAMI 2017 · similarity 0.89)
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.
