Research map / Image generation / Editing and style transfer
FaceDNeRF: Semantics-Driven Face Reconstruction, Prompt Editing and Relighting with Diffusion Models
Hao ZHANG, Tianyuan DAI, Yanbo Xu, Yu-Wing Tai, Chi-Keung Tang
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 |
|---|---|---|---|
| Use diffusion SDS guidance instead of CLIP guidance for text editing | FaceDNeRF · Text-edit examples vs StyleCLIP Latent Optimization | Clip loss / visual quality: 7.49 vs 6.53 | Mixed |
| Compare identity preservation against StyleCLIP Latent Optimization | FaceDNeRF · Text-edit examples | ID loss x100: 2.40 -> 2.26 | Improved |
| Edit in GAN latent space instead of InterfaceGAN linear direction (Young) | FaceDNeRF · Young edit | ID loss: 0.0159 -> 0.0102 | Improved |
| Edit in GAN latent space instead of InterfaceGAN linear direction (Smile) | FaceDNeRF · Smile edit | ID loss: 0.0105 -> 0.0028 | Improved |
| Edit in GAN latent space instead of InterfaceGAN linear direction (Mustache) | FaceDNeRF · Mustache edit | ID loss: 0.0148 -> 0.0014 | Improved |
| Increase identity loss weight lambda_ID (ablation) | FaceDNeRF · Prompt 'A bald man with a thick mustache' | similarity to input identity (qualitative) | Improved |
| Increase reconstruction loss weight lambda_R (ablation) | FaceDNeRF · Prompt 'A bald man with a thick mustache' | text-edit fidelity (qualitative) | Mixed |
| Increase diffusion loss weight lambda_D (ablation) | FaceDNeRF · Prompt 'A bald man with a thick mustache' | text guidance strength (qualitative) | Improved |
| Use SDS in GAN latent space rather than original SDS on unconstrained NeRF (ablation) | FaceDNeRF · Text-conditioned head generation | generation quality (qualitative) | Improved |
| Use SDS guidance instead of CLIP guidance (ablation) | FaceDNeRF · Face editing with attribute prompts | image quality (qualitative) | Improved |
| Use short difference-only prompt instead of full description (ablation) | FaceDNeRF · Editing prompts | edit success (qualitative) | No clear effect |
| Use ambiguous single-word prompt 'Old' or 'senior' instead of 'An old man' (ablation) | FaceDNeRF · Editing prompts | edit success (qualitative) | 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.
- InterfaceGAN — InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs (TPAMI 2020) · better on 3 of 3 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
- StrokeFaceNeRF: Stroke-based Facial Appearance Editing in Neural Radiance Field (CVPR 2024 · similarity 0.90)
- FFaceNeRF: Few-shot Face Editing in Neural Radiance Fields (CVPR 2025 · similarity 0.89)
- InstructPix2NeRF: Instructed 3D Portrait Editing from a Single Image (ICLR 2024 · similarity 0.88)
- OSDFace: One-Step Diffusion Model for Face Restoration (CVPR 2025 · similarity 0.87)
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.87)
- 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.
