Research map / Image generation / Editing and style transfer
Disentangling Writer and Character Styles for Handwriting Generation
Gang Dai, Yifan Zhang, Qingfeng Wang, Qing Du, Zhuliang Yu, Zhuoman Liu, Shuangping Huang
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 |
|---|---|---|---|
| disentangle writer-wise and character-wise style representations and train them with complementary contrastive objectives (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 85.52 -> 94.50 | Improved |
| add writer-wise style representation (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 85.52 -> 91.38 | Improved |
| add character-wise style representation (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 85.52 -> 90.31 | Improved |
| use lower GlyphNCE patch sampling ratio (0.25 rather than 0.50, 0.75, or 1.00) (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 92.07, 91.91, 91.54 -> 94.50 | Improved |
| replace SDT content/previous-point combination with concatenation (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 94.50 -> 91.61 | Hurt |
| replace attention-based content/style fusion with concatenation (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 94.50 -> 78.12 | Hurt |
| reverse the order of the two style features in the Transformer decoder (ablation) | SDT · ICDAR-2013 Chinese | Style Score: 93.72% vs. 94.50% | No clear effect |
| use content-style-conditioned generation instead of content-only generation (ablation) | SDT · Indic (Tamil) | Content Score: 68.50% -> 97.22% | Improved |
Closest papers by meaning
- Hierarchical Style Aggregation for Versatile Chinese Handwriting Generation (ECCV 2026 · similarity 0.91)
- Elegantly Written: Disentangling Writer and Character Styles for Enhancing Online Chinese Handwriting (ECCV 2024 · similarity 0.90)
- Learning to Generate Stylized Handwritten Text via a Unified Representation of Style, Content, and Noise (ICLR 2026 · similarity 0.90)
- DiffInk: Glyph- and Style-Aware Latent Diffusion Transformer for Text to Online Handwriting Generation (ICLR 2026 · similarity 0.89)
- Decoupling Layout from Glyph in Online Chinese Handwriting Generation (ICLR 2025 · similarity 0.89)
- A Multi-Style Chinese Characters Writing Intelligent Tool Based on Small-scale Training Data (AAAI 2025 · 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.
