Research map / 3D vision / NeRF and radiance fields
Learning 3D-Aware Image Synthesis With Unknown Pose Distribution
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 |
|---|---|---|---|
| Use PoF3D without pose priors instead of CAMPARI | 3D-aware image synthesis · FFHQ, Cats, Shapenet Cars | FID: 58.59 -> 4.99 (FFHQ); 37.40 -> 5.46 (Cats); 68.91 -> 3.78 (Shapenet Cars) | Helped |
| Add symmetry loss (ablation) | PoF3D · FFHQ | Depth Error: 0.41 -> 0.29 | Helped |
| Add a pose-aware discriminator (ablation) | PoF3D · FFHQ | Depth Error: 1.30 -> 0.29 | Helped |
| Condition discriminator classification on predicted pose (ablation) | PoF3D · FFHQ | Depth Error: 1.48 -> 0.29 | Helped |
| Increase pose learner learning rate to 2.5e-4 (ablation) | PoF3D · FFHQ | FID: 4.99 -> 122.07 | Hurt |
| Decrease pose learner learning rate to 2.5e-6 (ablation) | PoF3D · FFHQ | FID: 4.99 -> 10.20 | Hurt |
| Use pose-free generation with a learned pose from the latent code | PoF3D · Shapenet Cars | FID: 9.68* -> 3.78 | Helped |
Closest papers by meaning
- Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis (CVPR 2020 · similarity 0.90)
- PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis (ICML 2025 · similarity 0.86)
- Sparse-view Pose Estimation and Reconstruction via Analysis by Generative Synthesis (NeurIPS 2024 · similarity 0.86)
- Free3D: Consistent Novel View Synthesis without 3D Representation (CVPR 2024 · similarity 0.86)
- Source-Free and Image-Only Unsupervised Domain Adaptation for Category Level Object Pose Estimation (ICLR 2024 · similarity 0.86)
- Towards Robust 3D Pose Transfer with Adversarial Learning (CVPR 2024 · 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.