Research map / 3D vision / Face and shape
Pix3D: Dataset and Methods for Single-Image 3D Shape Modeling
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 |
|---|---|---|---|
| add joint pose estimation with the view estimator (ablation) | Ours on Pix3D chairs · Pix3D chairs | IoU: 0.267 -> 0.282 | Helped |
| add joint pose estimation with the view estimator (ablation) | Ours on Pix3D chairs · Pix3D chairs | EMD: 0.124 -> 0.118 | Helped |
| add joint pose estimation with the view estimator (ablation) | Ours on Pix3D chairs · Pix3D chairs | CD: 0.124 -> 0.119 | Helped |
| remove the pose estimation module (ablation) | Ours image-based retrieval · Pix3D image-based retrieval | R@1: 0.42 -> 0.53 | Helped |
| add joint pose estimation (ablation) | Ours image-based retrieval · Pix3D image-based retrieval | R@1: 0.53 -> 0.42 | Hurt |
| use the proposed joint shape and pose model instead of Render for CNN | Pix3D pose estimation · Pix3D pose estimation, 4 views | Azimuth accuracy: 0.71 -> 0.76 | Helped |
| use the proposed joint shape and pose model instead of Render for CNN | Pix3D pose estimation · Pix3D pose estimation, 4 views | Elevation accuracy: 0.57 -> 0.87 | Helped |
| evaluate reconstruction with Chamfer distance or Earth Mover's distance rather than IoU | 3D shape reconstruction metric evaluation · Human ratings of reconstructed shapes | Pearson correlation: 0.371, 0.544, and 0.518 | Helped |
Closest papers by meaning
- SPAR3D: Stable Point-Aware Reconstruction of 3D Objects from Single Images (CVPR 2025 · similarity 0.87)
- PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis (ICML 2025 · similarity 0.86)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.86)
- PSHuman: Photorealistic Single-image 3D Human Reconstruction using Cross-Scale Multiview Diffusion and Explicit Remeshing (CVPR 2025 · similarity 0.85)
- Object-X: Learning to Reconstruct Multi-Modal 3D Object Representations (NeurIPS 2025 · similarity 0.85)
- Tango3D: Towards Alignment for Global and Local 2D-3D Correspondence (NeurIPS 2026 · similarity 0.85)
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.