VecFontSDF: Learning to Reconstruct and Synthesize High-quality Vector Fonts via Signed Distance Functions

Teaser figure from the VecFontSDF paper. — VecFontSDF reconstructs and synthesizes high-quality vector glyphs through an SDF-based implicit representation that can be converted into quadratic Bézier curves for editable vector font generation.

Abstract

Font design is of vital importance in digital content design and the modern printing industry. Developing algorithms capable of automatically synthesizing vector fonts can significantly facilitate the font design process. However, existing methods mainly concentrate on raster image generation, and only a few approaches can directly synthesize vector fonts. This paper proposes an end-to-end trainable method, VecFontSDF, to reconstruct and synthesize high-quality vector fonts using signed distance functions (SDFs). Specifically, based on the proposed SDF-based implicit shape representation, VecFontSDF learns to model each glyph as shape primitives enclosed by several parabolic curves, which can be precisely converted to quadratic Bézier curves that are widely used in vector font products. In this manner, most image generation methods can be easily extended to synthesize vector fonts. Qualitative and quantitative experiments conducted on a publicly available dataset demonstrate that our method obtains high-quality results on several tasks, including vector font reconstruction, interpolation, and few-shot vector font synthesis, markedly outperforming the state of the art.

Method

Overview of the VecFontSDF framework. — Overview of VecFontSDF. The method represents glyph shapes with SDF-based implicit primitives and converts them into quadratic Bézier curves, enabling high-quality reconstruction and synthesis of vector fonts.

Results

Reconstruction

Interpolation

Few-shot Generation

Few-shot vector font generation results.

BibTeX

@InProceedings{Xia_2023_CVPR,
  author    = {Xia, Zeqing and Xiong, Bojun and Lian, Zhouhui},
  title     = {VecFontSDF: Learning To Reconstruct and Synthesize High-Quality Vector Fonts via Signed Distance Functions},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month     = {June},
  year      = {2023},
  pages     = {1848-1857}
}