Research map / 3D vision / NeRF and radiance fields
DeformToon3D: Deformable Neural Radiance Fields for 3D Toonification
Junzhe Zhang, Yushi Lan, Shuai Yang, Fangzhou Hong, Quan Wang, Chai Kiat Yeo, Ziwei Liu, Chen Change Loy
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 instance code wR as a condition of the StyleField (ablation) | DeformToon3D · Toonification, 10 styles | FID: 33.4 -> 27.6 | Improved |
| add instance code wR as a condition of the StyleField (ablation) | DeformToon3D · Toonification, 10 styles | Identity preservation: 0.735 -> 0.781 | Improved |
| use SIREN layers instead of a plain MLP for the StyleField (ablation) | DeformToon3D · Toonification, 10 styles | FID: 31.8 -> 27.6 | Improved |
| add adversarial loss (ablation) | DeformToon3D · Toonification, 10 styles | FID: 29.9 -> 27.6 | Improved |
| add adversarial loss (ablation) | DeformToon3D · Toonification, 10 styles | Identity preservation: 0.769 -> 0.781 | Improved |
| keep 3D generator frozen and deform with StyleField instead of fine-tuning the whole generator | 3D toonification (StyleSDF) · Toonification, 10 styles | Identity preservation: 0.707 -> 0.781 | Improved |
| fine-tune both G0 and G1 rather than only the super-resolution G1 | CIPS-3D vs E3DGE baselines · Toonification, 10 styles | FID: 50.6 -> 34.0 | Improved |
| use a single multi-style StyleField instead of per-style fine-tuned models | DeformToon3D · Model storage, 10 styles | storage saving: 98.5% | Improved |
| apply directional CLIP loss supervision to a 3D GAN (StyleGAN-NADA extension) | 3D GAN toonification · Toonification, 10 styles | FID: 59.3 | Hurt |
Later papers that compared with this method
1 later accepted paper reports DeformToon3D as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- 3DToonify: Creating Your High-Fidelity 3D Stylized Avatar Easily from 2D Portrait Images (CVPR 2024 · better on 2 of 2 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
- Towards High-Quality and Disentangled Face Editing in a 3D GAN (TPAMI 2025 · similarity 0.89)
- Next3D: Generative Neural Texture Rasterization for 3D-Aware Head Avatars (CVPR 2023 · similarity 0.89)
- 3DToonify: Creating Your High-Fidelity 3D Stylized Avatar Easily from 2D Portrait Images (CVPR 2024 · similarity 0.87)
- In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 2024 · similarity 0.87)
- FPRF: Feed-Forward Photorealistic Style Transfer of Large-Scale 3D Neural Radiance Fields (AAAI 2024 · similarity 0.87)
- LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example (CVPR 2024 · similarity 0.86)
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.
