Research map / 3D vision / Face and shape
2D GANs Meet Unsupervised Single-View 3D Reconstruction
Feng Liu, Xiaoming Liu
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 uncertainty prediction module (ablation) | Proposed · PASCAL3D+ car | Chamfer distance: 0.208 -> 0.195 | Improved |
| add uncertainty prediction module (ablation) | Proposed · PASCAL3D+ car | 3D IoU: 0.681 -> 0.702 | Improved |
| increase pseudo views from n=10 to n=40 (ablation) | Proposed · PASCAL3D+ car | Chamfer distance: 0.336 -> 0.195 | Improved |
| increase pseudo views from n=20 to n=40 (ablation) | Proposed · PASCAL3D+ car | 3D IoU: 0.643 -> 0.702 | Improved |
| increase pseudo views from n=40 to n=50 (ablation) | Proposed · PASCAL3D+ car | 3D IoU: 0.702 -> 0.714 | No clear effect |
| increase pseudo views from n=40 to n=50 (ablation) | Proposed · PASCAL3D+ car | Chamfer distance: 0.195 -> 0.191 | No clear effect |
| use proposed eigenvector-filtered multi-view generation instead of Baseline swap (ablation) | Proposed · PASCAL3D+ car | 3D IoU: 0.679 -> 0.702 | Improved |
| train SDF-SRN on GAN-generated pseudo images | SDF-SRN · PASCAL3D+ car | 3D IoU: 0.653 -> 0.661 | Improved |
| train SDF-SRN on GAN-generated pseudo images | SDF-SRN · PASCAL3D+ airplane | 3D IoU: 0.405 -> 0.412 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SDF-SRN — SDF-SRN: Learning Signed Distance 3D Object Reconstruction from Static Images (NeurIPS 2020) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Progressive Learning of 3D Reconstruction Network From 2D GAN Data (TPAMI 2023 · similarity 0.90)
- Image GANs meet Differentiable Rendering for Inverse Graphics and Interpretable 3D Neural Rendering (ICLR 2021 · similarity 0.89)
- Efficient Geometry-Aware 3D Generative Adversarial Networks (CVPR 2022 · similarity 0.89)
- High-Fidelity 3D GAN Inversion by Pseudo-Multi-View Optimization (CVPR 2023 · similarity 0.89)
- What You See is What You GAN: Rendering Every Pixel for High-Fidelity Geometry in 3D GANs (CVPR 2024 · similarity 0.87)
- Dual Encoder GAN Inversion for High-Fidelity 3D Head Reconstruction from Single Images (NeurIPS 2024 · 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.
