Research map / Image generation / Editing and style transfer
VecDesigner: Exploring Visual Guidance and Structural Consistency for Semantic Typography
Liu Yu, Xingjiao Wu, Ziang Liu, Jiabao Zhao, Daoguo Dong, Liang He
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 VGSDS visual guidance to vanilla SDS (ablation) | vanilla SDS baseline · Semantic typography (100 concepts, 5 fonts) | OCR accuracy (%): 81.26 -> 83.55 | Improved |
| add Procrustes loss to VGSDS (ablation) | VGSDS · Semantic typography (100 concepts, 5 fonts) | Chamfer distance: 1.4816 -> 1.4495 | Improved |
| add intersection loss to VGSDS + Procrustes (ablation) | VGSDS + Procrustes · Semantic typography (100 concepts, 5 fonts) | Chamfer distance: 1.4495 -> 1.3717 | Improved |
| add VGSDS to Procrustes + intersection loss (ablation) | VecDesigner · Semantic typography (100 concepts, 5 fonts) | OCR accuracy (%): 84.33 -> 85.36 | Improved |
| add Procrustes loss to VGSDS + intersection loss (ablation) | VecDesigner · Semantic typography (100 concepts, 5 fonts) | BLIPScore: 0.3467 -> 0.3535 | Improved |
| update reference images every 500 steps instead of 50 (ablation) | VecDesigner · Semantic typography (100 concepts, 5 fonts) | OCR accuracy (%): 83.21 -> 85.36 | Improved |
| update reference images every 500 steps instead of 100 (ablation) | VecDesigner · Semantic typography (100 concepts, 5 fonts) | Chamfer distance: 1.4455 -> 1.3717 | Improved |
| use 10 additional sampling points instead of none (ablation) | Procrustes loss · Semantic typography (100 concepts, 5 fonts) | OCR accuracy (%): 84.08 -> 85.36 | Improved |
| use 10 additional sampling points instead of 20 (ablation) | Procrustes loss · Semantic typography (100 concepts, 5 fonts) | CLIPScore: 0.2733 -> 0.2688 | Hurt |
| vary lambda_v, lambda_p, lambda_s (ablation) | VecDesigner · Semantic typography (100 concepts, 5 fonts) | all metrics | No clear effect |
| use VecDesigner instead of Word-As-Image | vector-based semantic typography · Semantic typography (100 concepts, 5 fonts) | CLIPScore: 0.2518 -> 0.2733 | Improved |
| use VecDesigner instead of Word-As-Image | vector-based semantic typography · Semantic typography (100 concepts, 5 fonts) | OCR accuracy (%): 82.36 -> 85.36 | 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.
- DS-Fusion — DS-Fusion: Artistic Typography via Discriminated and Stylized Diffusion (ICCV 2023) · better on 3 of 3 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
- VecGlypher: Unified Vector Glyph Generation with Language Models (CVPR 2026 · similarity 0.89)
- MetaDesigner: Advancing Artistic Typography through AI-Driven, User-Centric, and Multilingual WordArt Synthesis (ICLR 2025 · similarity 0.86)
- Towards Visual Text Design Transfer Across Languages (NeurIPS 2024 · similarity 0.86)
- Type-R: Automatically Retouching Typos for Text-to-Image Generation (CVPR 2025 · similarity 0.85)
- DesignDiffusion: High-Quality Text-to-Design Image Generation with Diffusion Models (CVPR 2025 · similarity 0.85)
- ASCIIEval: Benchmarking Models' Visual Perception in Text Strings via ASCII Art (ICLR 2026 · 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.
