Research map / Image generation / Text-to-image diffusion
SkyReels-Text: Fine-Grained Font-Controllable Text Editing for Poster Design
Yunjie Yu, Jingchen Wu, Junchen Zhu, Chunze Lin, Guibin Chen
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 font style reference glyph map as second condition (ablation) | SkyReels-Text · SkyReels-Text Benchmark | DINO: 0.6995 -> 0.8503 | Improved |
| add font style reference glyph map as second condition (ablation) | SkyReels-Text · SkyReels-Text Benchmark | Sen. Acc: 0.9327 -> 0.8334 | Hurt |
| add font style reference glyph map as second condition (ablation) | SkyReels-Text · SkyReels-Text Benchmark | Spatial IoU: 0.6986 -> 0.7506 | Improved |
| add text-region weighted loss (lambda 0 -> 5) (ablation) | SkyReels-Text · SkyReels-Text Benchmark | Sen. Acc: 0.7998 -> 0.8334 | Improved |
| increase loss weight lambda from 5 to 10 (ablation) | SkyReels-Text · SkyReels-Text Benchmark | B-PSNR: Sen. Acc 0.8334 -> 0.8363; B-PSNR 34.17 -> 33.13 | Mixed |
| SkyReels-Text vs FLUX-Text on background preservation | SkyReels-Text · SkyReels-Text Benchmark | B-PSNR: 34.53 -> 34.17 | Hurt |
| SkyReels-Text vs Qwen-Image-Edit baseline on AnyText English | Qwen-Image-Edit · AnyText benchmark (English) | FID: 35.62 -> 6.12 | Improved |
| SkyReels-Text vs Qwen-Image-Edit baseline on AnyText Chinese | Qwen-Image-Edit · AnyText benchmark (Chinese) | Sen. Acc: 0.6735 -> 0.7710 | Improved |
| zero-shot handwriting generation with single reference | SkyReels-Text · CVL | HWD: 1.06 -> 0.89 | Improved |
| zero-shot handwriting generation, geometry score | SkyReels-Text · CVL | GS: 29.6 -> 31.0 | Hurt |
Closest papers by meaning
- VecGlypher: Unified Vector Glyph Generation with Language Models (CVPR 2026 · similarity 0.85)
- GlyphDraw2: Automatic Generation of Complex Glyph Posters with Diffusion Models and Large Language Models (AAAI 2025 · similarity 0.84)
- Type-R: Automatically Retouching Typos for Text-to-Image Generation (CVPR 2025 · similarity 0.84)
- TextCtrl: Diffusion-based Scene Text Editing with Prior Guidance Control (NeurIPS 2024 · similarity 0.83)
- Towards Visual Text Design Transfer Across Languages (NeurIPS 2024 · similarity 0.83)
- MetaDesigner: Advancing Artistic Typography through AI-Driven, User-Centric, and Multilingual WordArt Synthesis (ICLR 2025 · similarity 0.83)
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.
