Calligrapher: Freestyle Text Image Customization
Yue Ma, Qingyan Bai, Hao Ouyang, Ka Leong Cheng, Qiuyu Wang, Hongyu Liu, Zichen Liu, Haofan Wang, Jingye Chen, Yujun Shen, Qifeng Chen
2025-07-01
Summary
This paper talks about Calligrapher, a new AI system that allows users to customize text images with artistic styles, creating beautiful typography in different fonts and designs.
What's the problem?
Traditional text image tools are often limited in how much users can customize the look of letters, making it hard to create unique and visually consistent artistic text designs.
What's the solution?
Calligrapher uses a diffusion-based AI model combined with techniques like self-distillation that help the model learn and improve itself, a special localized style injection that applies artistic effects precisely to parts of the text, and in-context generation to keep the style consistent throughout the text image.
Why it matters?
This matters because it gives artists, designers, and everyday users powerful tools to create customized, stylish text images quickly and easily, opening new possibilities for graphic design and creative expression.
Abstract
Calligrapher, a diffusion-based framework, integrates advanced text customization with artistic typography using self-distillation, a localized style injection framework, and in-context generation to achieve high-quality, visually consistent typography.