Research map / Image generation / Text-to-image diffusion
FreeText: Training-Free Text Rendering via Attention Localization and Spectral Glyph Injection
RuiQiang Zhang, Hengyi Wang, Chang Liu, Guanjie Wang, Zehua Ma, Weiming Zhang
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 FreeText | Qwen-Image · longText-en | NED: 0.625 -> 0.713 | Improved |
| add FreeText | FLUX.1-dev · longText-en | VQA Score: 2.563 -> 4.211 | Improved |
| add FreeText | Qwen-Image · longText-zh | NED: 0.639 -> 0.694 | Improved |
| add FreeText | FLUX.1-dev · CVTG | NED: 0.712 -> 0.722 | Improved |
| add FreeText | SD3-M · CVTG | NED: 0.616 -> 0.669 | Improved |
| add FreeText | SD3.5-L · CVTG | CLIPScore: 0.879 -> 0.871 | No clear effect |
| add FreeText | FLUX.1-dev · longText-en | AestheticScore: 5.365 -> 5.342 | No clear effect |
| add FreeText | Qwen-Image · CLT-Bench | NED: 0.458 -> 0.488 | Improved |
| add SGMI spectral modulation on top of FreeText localization (ablation) | Qwen-Image · longText-en | NED: 0.686 -> 0.713 | Improved |
| add SGMI spectral modulation (ablation) | FLUX.1-dev · longText-en | VQA Score: 3.816 -> 4.211 | Improved |
| use entity plus sink tokens instead of entity only for localization (ablation) | FreeText · text localization | IoU: 0.495 -> 0.561 | Improved |
| use sink-only tokens instead of entity + sink (ablation) | FreeText · text localization | IoU: 0.561 -> 0.479 | Hurt |
Closest papers by meaning
- GlyphMastero: A Glyph Encoder for High-Fidelity Scene Text Editing (CVPR 2025 · similarity 0.90)
- GlyphAnchor: Enhancing Visual Text Rendering via Position-Anchored Glyph Priors (NeurIPS 2026 · similarity 0.89)
- DreamText: High Fidelity Scene Text Synthesis (CVPR 2025 · similarity 0.88)
- Precise Parameter Localization for Textual Generation in Diffusion Models (ICLR 2025 · similarity 0.88)
- TextCtrl: Diffusion-based Scene Text Editing with Prior Guidance Control (NeurIPS 2024 · similarity 0.88)
- Magnifying What Matters: Attention-Guided Adaptive Rendering for Visual Text Comprehension (NeurIPS 2026 · 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.
