Research map / Image generation / Autoregressive and latent generation
VecFontSDF: Learning To Reconstruct and Synthesize High-Quality Vector Fonts via Signed Distance Functions
Zeqing Xia, Bojun Xiong, Zhouhui Lian
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 grid SDF and contour SDF supervision (vs. image loss only) (ablation) | VecFontSDF reconstruction · Vector font reconstruction (test set) | PSNR: 17.7928 -> 27.8502 (full model) | Improved |
| add contour SDF loss on top of grid SDF loss (ablation) | VecFontSDF reconstruction · Vector font reconstruction (test set) | L1 distance: 0.0110 -> 0.0090 | Improved |
| add contour SDF loss on top of grid SDF loss (ablation) | VecFontSDF reconstruction · Vector font reconstruction (test set) | LPIPS: 0.0384 -> 0.0303 | Improved |
| add grid and contour SDF supervision (vs. image loss only) (ablation) | VecFontSDF reconstruction · Vector font reconstruction (test set) | IoU: 0.9569 -> 0.9901 | Improved |
| VecFontSDF vs. IGSR and BSP-Net at 64x64 resolution | Implicit shape representation · Vector font reconstruction (64x64) | PSNR: 24.6895 (IGSR) -> 28.6739 | Improved |
| Few-shot style transfer with VecFontSDF decoder vs DeepVecFont | Few-shot vector font synthesis · Qualitative comparison (Fig. 13) | qualitative shape quality | Improved |
| Removing SDF-supervised losses in favor of image-only loss (ablation) | VecFontSDF reconstruction · Vector font reconstruction (test set) | qualitative shape fidelity | Hurt |
Closest papers by meaning
- DeepVecFont-v2: Exploiting Transformers To Synthesize Vector Fonts With Higher Quality (CVPR 2023 · similarity 0.91)
- VecFusion: Vector Font Generation with Diffusion (CVPR 2024 · similarity 0.88)
- VecGlypher: Unified Vector Glyph Generation with Language Models (CVPR 2026 · similarity 0.87)
- VQ-FONT: Few-Shot Font Generation with Structure-Aware Enhancement and Quantization (AAAI 2024 · similarity 0.85)
- Generate Like Experts: Multi-Stage Font Generation by Incorporating Font Transfer Process into Diffusion Models (CVPR 2024 · similarity 0.84)
- DeepCalliFont: Few-Shot Chinese Calligraphy Font Synthesis by Integrating Dual-Modality Generative Models (AAAI 2024 · similarity 0.84)
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.
