Research map / 3D vision / NeRF and radiance fields
Generative Neural Articulated Radiance Fields
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 |
|---|---|---|---|
| train generator in canonical pose with explicit SF deformation vs EG3D without warping (ablation) | GNARF · SURREAL | FID (10k): 14.3 -> 4.7 | Helped |
| train generator in canonical pose with explicit SF deformation vs EG3D without warping (ablation) | GNARF · AIST++ | FID (50k): 8.3 -> 7.9 | Helped |
| deformation-aware training instead of post-hoc pose estimation and re-warping of EG3D (ablation) | GNARF · AIST++ | PCKh@0.5: 0.855 -> 0.980 | Helped |
| deformation-aware training instead of post-hoc re-warping of EG3D (ablation) | GNARF · SURREAL | PCKh@0.5: 0.348 -> 0.999 | Helped |
| deformation-aware training instead of post-hoc re-warping of EG3D (ablation) | GNARF · FFHQ | FID (50k): 11.6 -> 6.6 | Helped |
| deformation-aware training vs EG3D re-warping on identity consistency (ablation) | GNARF · FFHQ | ID-Consistency: 0.81 -> 0.80 | No clear effect |
| Surface Field deformation instead of MVC (ablation) | single-scene tri-plane · AIST++ single-scene overfitting | PSNR (test): 17.2 -> 18.0 | Helped |
| Surface Field deformation instead of skinning (ablation) | single-scene tri-plane · AIST++ single-scene overfitting | PSNR (test): 17.9 -> 18.0 | No clear effect |
| Surface Field deformation instead of skinning (ablation) | single-scene tri-plane · AIST++ single-scene overfitting | run time (ms): 95.6 -> 31.6 | Helped |
Closest papers by meaning
- A-NeRF: Articulated Neural Radiance Fields for Learning Human Shape, Appearance, and Pose (NeurIPS 2021 · similarity 0.91)
- Pix2NeRF: Unsupervised Conditional p-GAN for Single Image to Neural Radiance Fields Translation (CVPR 2022 · similarity 0.89)
- HumanGen: Generating Human Radiance Fields With Explicit Priors (CVPR 2023 · similarity 0.89)
- GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis (NeurIPS 2020 · similarity 0.89)
- H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion (NeurIPS 2021 · similarity 0.89)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 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.