Research map / 3D vision / 3D generation
MaTe3D: Mask-Guided Text-based 3D-Aware Portrait Editing
Kangneng Zhou, Daiheng Gao, Xu An Wang, Jie Zhang, Peng Zhang, Xusen Sun, Longhao Zhang, Shiqi Yang, Bang Zhang, Liefeng Bo, Yaxing Wang, Ming-Ming Cheng
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 dual tri-planes generation (ablation) | MaTe3D generator · FFHQ generation | FID: 5.8 -> 5.1 | Improved |
| add dual tri-planes generation (ablation) | MaTe3D generator · FFHQ generation | KID: 0.389 -> 0.215 | Improved |
| add dual tri-planes generation (ablation) | MaTe3D generator · AFHQ cat generation | FID: 3.9 -> 2.5 | Improved |
| add dual tri-planes generation (ablation) | MaTe3D generator · FFHQ generation | multi-view ID consistency: 0.75 -> 0.79 | Improved |
| add global and local SDF representations (ablation) | MaTe3D generator · Mask accuracy, 1,000 sampled FFHQ images | mIoU: 0.812 -> 0.873 | Improved |
| add density consistency loss (ablation) | MaTe3D generator · FFHQ generation | image quality (qualitative) | Improved |
| add SDF consistency loss (ablation) | MaTe3D generator · FFHQ generation | mask aliasing and convergence (qualitative) | Improved |
| add Condition Updating Scheme in CDGT (ablation) | MaTe3D editing · FFHQ editing | texture stability (qualitative) | Improved |
| add gradient combination of image and normal map (ablation) | MaTe3D editing · FFHQ editing | novel-view geometry (qualitative) | Improved |
| replace constant TADA blending with random blending of image and normal gradients (ablation) | MaTe3D editing · FFHQ editing | texture and geometry quality (qualitative) | Improved |
| use MaTe3D local SDFs instead of IDE-3D for part decomposition | MaTe3D vs IDE-3D · Decompositional geometry, parts: nose, mouth, hair, ear | Chamfer-L1: 2677.9805 -> 528.7918 | Improved |
| use MaTe3D instead of IDE-3D+StyleCLIP for text editing | MaTe3D vs IDE-3D+StyleCLIP · FFHQ text-based editing | CLIPScore: 0.762 -> 0.791 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- BallGAN — BallGAN: 3D-aware Image Synthesis with a Spherical Background (ICCV 2023) · better on 3 of 3 shared results
- FENeRF — FENeRF: Face Editing in Neural Radiance Fields (CVPR 2022) · better on 1 of 1 shared result
- StyleSDF — StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation (CVPR 2022) · better on 1 of 1 shared result
- StyleNeRF — StyleNeRF: A Style-based 3D Aware Generator for High-resolution Image Synthesis (ICLR 2022) · better on 1 of 1 shared result
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
- Text-Guided 3D Face Synthesis - From Generation to Editing (CVPR 2024 · similarity 0.88)
- InstructPix2NeRF: Instructed 3D Portrait Editing from a Single Image (ICLR 2024 · similarity 0.88)
- Fast Text-to-3D-Aware Face Generation and Manipulation via Direct Cross-modal Mapping and Geometric Regularization (ICML 2024 · similarity 0.88)
- 3D-Aware Face Editing via Warping-Guided Latent Direction Learning (CVPR 2024 · similarity 0.87)
- StrokeFaceNeRF: Stroke-based Facial Appearance Editing in Neural Radiance Field (CVPR 2024 · similarity 0.87)
- FFaceNeRF: Few-shot Face Editing in Neural Radiance Fields (CVPR 2025 · 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.
