Research map / 3D vision / NeRF and radiance fields
Injecting 3D Perception of Controllable NeRF-GAN into StyleGAN for Editable Portrait Image Synthesis
Jeong-gi Kwak, Yuanming Li, Dongsik Yoon, Donghyeon Kim, David Han, Hanseok Ko
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 pose loss to SURF-GAN training (ablation) | SURF-GAN · FFHQ 128x128 | pose error (x10^-2): 2.69 -> 2.36 | Improved |
| add pose loss to SURF-GAN training (ablation) | SURF-GAN · FFHQ 128x128 | FID: 44.56 -> 45.92 | Hurt |
| add progressive growing to SURF-GAN training (ablation) | SURF-GAN · CelebA 64x64 | FID | Improved |
| SURF-GAN vs pi-GAN on FFHQ image quality | SURF-GAN · FFHQ 64x64 trained | FID: 47.12 -> 44.56 | Improved |
| SURF-GAN vs pi-GAN on CelebA pose accuracy | SURF-GAN · CelebA 64x64 trained | pose error (x10^-2): 5.81 -> 6.10 | Hurt |
| SURF-GAN vs pi-GAN on multi-view identity consistency | SURF-GAN · CelebA 64x64 trained | identity similarity: 0.65 -> 0.68 | Improved |
| inject SURF-GAN 3D control into StyleGAN2 vs CIPS-3D | 3D-controllable StyleGAN · FFHQ 256x256 | FID: 6.97 -> 4.72 | Improved |
| inject SURF-GAN 3D control into StyleGAN2 vs pi-GAN speed | 3D-controllable StyleGAN · FFHQ 256x256 | frames per second: 4 -> 72 | Improved |
| 3D-controllable StyleGAN vs pi-GAN on pose accuracy | 3D-controllable StyleGAN · FFHQ 256x256 | pose error (x10^-2): 3.81 -> 4.24 | Hurt |
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.
- pi-GAN — Pi-GAN: Periodic Implicit Generative Adversarial Networks for 3D-Aware Image Synthesis (CVPR 2021) · better on 7 of 10 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
- Dual Mapping of 2D StyleGAN for 3D-Aware Image Generation and Manipulation (Student Abstract) (AAAI 2024 · similarity 0.92)
- Lifting 2D StyleGAN for 3D-Aware Face Generation (CVPR 2021 · similarity 0.92)
- MOST-GAN: 3D Morphable StyleGAN for Disentangled Face Image Manipulation (AAAI 2022 · similarity 0.90)
- 3D-FM GAN: Towards 3D-Controllable Face Manipulation (ECCV 2022 · similarity 0.90)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 2024 · similarity 0.89)
- Reference-Based 3D-Aware Image Editing with Triplanes (CVPR 2025 · similarity 0.88)
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.
