Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

WordArt Designer API: User-Driven Artistic Typography Synthesis with Large Language Models on ModelScope

Jun-Yan He, Zhi-Qi Cheng, Chenyang Li, Jingdong Sun, Wangmeng Xiang, Yusen Hu, Xianhui Lin, Xiaoyang Kang, Zengke Jin, Bin Luo, Yifeng Geng, Xuansong Xie, Jingren Zhou

TL;DR

The paper addresses barriers to artistic typography by proposing a user-driven, LLM-powered design pipeline that maps free-form input to SemTypo, StyTypo, and TexTypo transformations on ModelScope. The approach employs an iterative quality loop with a threshold of $K$ successful transformations, yielding four design variations within roughly one minute per run. Key contributions include a four-service WordArt API, handwriting font generation, and real-world deployment with substantial user engagement and industry use (e.g., Alibaba), illustrating practical utility. This work advances personalized digital typography by empowering non-experts to craft aesthetically diverse typography while maintaining a feedback-driven design process.

Abstract

This paper introduces the WordArt Designer API, a novel framework for user-driven artistic typography synthesis utilizing Large Language Models (LLMs) on ModelScope. We address the challenge of simplifying artistic typography for non-professionals by offering a dynamic, adaptive, and computationally efficient alternative to traditional rigid templates. Our approach leverages the power of LLMs to understand and interpret user input, facilitating a more intuitive design process. We demonstrate through various case studies how users can articulate their aesthetic preferences and functional requirements, which the system then translates into unique and creative typographic designs. Our evaluations indicate significant improvements in user satisfaction, design flexibility, and creative expression over existing systems. The WordArt Designer API not only democratizes the art of typography but also opens up new possibilities for personalized digital communication and design.

WordArt Designer API: User-Driven Artistic Typography Synthesis with Large Language Models on ModelScope

TL;DR

The paper addresses barriers to artistic typography by proposing a user-driven, LLM-powered design pipeline that maps free-form input to SemTypo, StyTypo, and TexTypo transformations on ModelScope. The approach employs an iterative quality loop with a threshold of successful transformations, yielding four design variations within roughly one minute per run. Key contributions include a four-service WordArt API, handwriting font generation, and real-world deployment with substantial user engagement and industry use (e.g., Alibaba), illustrating practical utility. This work advances personalized digital typography by empowering non-experts to craft aesthetically diverse typography while maintaining a feedback-driven design process.

Abstract

This paper introduces the WordArt Designer API, a novel framework for user-driven artistic typography synthesis utilizing Large Language Models (LLMs) on ModelScope. We address the challenge of simplifying artistic typography for non-professionals by offering a dynamic, adaptive, and computationally efficient alternative to traditional rigid templates. Our approach leverages the power of LLMs to understand and interpret user input, facilitating a more intuitive design process. We demonstrate through various case studies how users can articulate their aesthetic preferences and functional requirements, which the system then translates into unique and creative typographic designs. Our evaluations indicate significant improvements in user satisfaction, design flexibility, and creative expression over existing systems. The WordArt Designer API not only democratizes the art of typography but also opens up new possibilities for personalized digital communication and design.
Paper Structure (6 sections, 5 figures)

This paper contains 6 sections, 5 figures.

Table of Contents

  1. Introduction
  2. Method
  3. Evaluation
  4. Ethical Implications
  5. System Framework Diagram
  6. Additional Examples of Fonts Generated by WordART Designer

Figures (5)

  • Figure 1: Annotated screenshot showcasing our system deployed on ModelScope. Four main services are provided: (1) A "one-stop shop" that directly outputs artistic transformations based on the user input. (2) and (3) Fine-grained control of intermediate results from the SemTypo, StyTypo, and TexTypo modules. (4) Personalized font generation using handwriting. You are welcome to explore more of WordART Designer at https://www.modelscope.cn/studios/WordArt/WordArt
  • Figure 2: WordART Designer's various components and how they interact with each other.
  • Figure 3: Typographies targeting fruits and foods.
  • Figure 4: Typographies targeting jewelry.
  • Figure 5: Typographies targeting plants and vegetables.