Few-Part-Shot Font Generation

2025-09-12

AI Summary

This work addresses the limitation in few-shot font generation that requires complete character inputs, proposing a novel paradigm for generating entire fonts from only a few local glyph components. Methodologically, we design a deep generative model that jointly performs component-level feature disentanglement and cross-character style propagation, enabling end-to-end inference from partial shapes to global glyph structure and consistent typographic style. Experiments on multiple standard font datasets demonstrate that our approach achieves high-fidelity generation of thousands of characters using merely 3–5 component glyphs as input—substantially outperforming baseline methods reliant on full-character exemplars. Our core contributions are threefold: (1) the first formalization of few-shot font generation as a “local-to-global” style generalization problem; (2) elimination of the strict requirement for complete character inputs, thereby lowering design barriers; and (3) empirical validation of the transferability of component-level features in governing holistic glyph structure.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

This paper proposes a novel model of few-part-shot font generation, which designs an entire font based on a set of partial design elements, i.e., partial shapes. Unlike conventional few-shot font generation, which requires entire character shapes for a couple of character classes, our approach only needs partial shapes as input. The proposed model not only improves the efficiency of font creation but also provides insights into how partial design details influence the entire structure of the individual characters.

Problem

• Generating complete fonts from partial design elements
• Improving efficiency of font creation process
• Analyzing impact of partial details on character structure

Innovation

• Generates fonts from partial design elements
• Uses few-shot learning with partial shapes
• Improves efficiency and provides structural insights