Research map / Image generation / Editing and style transfer
TexEditor: Structure-Preserving Text-Driven Texture Editing
Bo Zhao, Yihang Liu, Chenfeng Zhang, Huan Yang, Kun Gai, Wei Ji
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 |
|---|---|---|---|
| SFT on TexBlender scene-level data (vs none) (ablation) | Qwen-Image-Edit-2509 · TexBench texture replacement | TexEval: 0.717 -> 0.767 | Improved |
| use TexBlender instead of Alchemist single-object SFT data (ablation) | Qwen-Image-Edit-2509 · TexBench attribute editing | TexEval: 0.583 -> 0.620 | Improved |
| add structure-aware RL with normalized structure loss after TexBlender SFT (ablation) | TexEditor · TexBench texture replacement | TexEval: 0.767 -> 0.886 | Improved |
| add structure loss to instruction-only RL (ablation) | TexEditor · TexBench attribute editing | structure score: 0.473 -> 0.816 | Improved |
| add instruction loss to structure-only RL (ablation) | TexEditor · TexBench texture replacement | instruction score: 0.752 -> 0.858 | Improved |
| normalize structure score with task-specific thresholds (ablation) | TexEditor RL · TexBench attribute editing | TexEval: 0.690 -> 0.760 | Improved |
| use instruction-only RL loss (ablation) | TexEditor RL · TexBench | structure score: 0.723 -> 0.602 (texture) | Hurt |
| stronger cold-start SFT data before RL (ablation) | TexEditor RL · TexBench texture replacement | TexEval: 0.801 (no SFT) -> 0.831 (Alchemist) -> 0.886 (TexBlender) | Improved |
| use SSIM on SAUGE wireframes rather than wireframe IoU as structure distance (ablation) | StructureNFT · structure reward | robustness | Improved |
| set TexEval alpha to 0.6 vs 0.0 or 1.0 (ablation) | TexEval · human preference ranking consistency | ranking accuracy | Improved |
| fine-tune with TexEditor on texture data | Qwen-2509 · ImgEdit Action | score: 3.33 -> 2.79 | Hurt |
| fine-tune with TexEditor on texture data | Qwen-2509 · ImgEdit Background | score: 3.87 -> 4.13 | Improved |
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.
- Alchemist — Alchemist: Parametric Control of Material Properties with Diffusion Models (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
- TexTailor: Customized Text-aligned Texturing via Effective Resampling (ICLR 2025 · similarity 0.87)
- Text-Guided 3D Face Synthesis - From Generation to Editing (CVPR 2024 · similarity 0.86)
- TextCtrl: Diffusion-based Scene Text Editing with Prior Guidance Control (NeurIPS 2024 · similarity 0.86)
- ParallelEdits: Efficient Multi-Aspect Text-Driven Image Editing with Attention Grouping (NeurIPS 2024 · similarity 0.86)
- ViTeX-Bench: Benchmarking High-Fidelity Video Scene Text Editing (NeurIPS 2026 · similarity 0.86)
- Style-Editor: Text-driven Object-centric Style Editing (CVPR 2025 · similarity 0.85)
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.
