LoGAN enables multilingual font generation with few examples
LoGAN: Multilingual Font Localization with Generative Agents
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Creating fonts for many languages is hard because letters need matching style, spacing, and texture. The authors developed LoGAN, a tool that uses a few letters from a font or logo to produce full sets of characters for different languages, including complex ones like Chinese and Korean. LoGAN splits the task into small parts, such as generating letter shapes, adjusting style, spacing letters correctly, and expanding textures, coordinating all with an AI agent. Tests show LoGAN makes fonts that look accurate and consistent across languages better than existing methods.
What this means in practice
- •For font designers: Automatically expand font designs to cover multiple languages using very few example glyphs, preserving style and spacing.
- •For graphic design teams: Create consistent multilingual logo fonts by generating matching characters from limited examples in diverse writing systems.
Authors
Zhuoning Yuan, Ta-Ying Cheng, Benjamin Klein
Abstract
Localizing a font into new languages is a highly intricate task requiring precise design adaptation of glyphs, color/texture, and spacing/kerning, from source to target languages. Most existing methods focus on single glyph generation with limited capability in handling multilingual font rendering. In this work, we propose LoGAN, a VLM-based agentic framework for few-shot multilingual font localization, which takes in a small number of individual glyphs from a font or letters from a logo and uses them to generate complete character sets in other languages. LoGAN breaks down this task into multiple components: a glyph-level diffusion model, a style finetuning module, a spacing and kerning transfer algorithm, and a texture expansion model, with a VLM agent coordinator. LoGAN achieves broad language coverage for font localization with various styles, including Chinese/Korean/Japanese (CJK). We evaluate our approach on both font and real-world logo datasets spanning more than 27 languages and compare it against both specialized font generation and state-of-the-art image editing models with strong text rendering capabilities (e.g., FLUX, Nano-Banana). Our approach yields higher glyph fidelity while maintaining better style, texture, and kerning consistency according to both quantitative and qualitative evaluations.