FourCornerGAN: Glyph formation augmentation for unpaired Chinese font generation

IF 2.9 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Digital Signal Processing Pub Date : 2025-05-12 DOI:10.1016/j.dsp.2025.105305

Kaili Wang, Chi Zhou, Yuanlin Shi, Tianquan Wu, Chen Chen

{"title":"FourCornerGAN: Glyph formation augmentation for unpaired Chinese font generation","authors":"Kaili Wang, Chi Zhou, Yuanlin Shi, Tianquan Wu, Chen Chen","doi":"10.1016/j.dsp.2025.105305","DOIUrl":null,"url":null,"abstract":"<div><div>Chinese character font generation poses unique challenges due to the complexity of glyph structures and the scarcity of paired training data. Existing methods for Chinese character font generation often suffer from issues like missing glyph formation and insufficient detail. To overcome these limitations, combining with the spatial glyph formation information, a novel encoding method based on the Four-Corner Number is proposed and integrated into CycleGAN to develop into FourCornerGAN to enhance structural representation in unpaired Chinese font generation, and a new Four-Corner Consistency Loss is introduced to guide both the generator and discriminator in preserving spatial glyph formation details. Extensive experiments demonstrate that FourCornerGAN significantly improves generation quality over baseline models, particularly in structural accuracy and visual consistency. This approach offers a promising solution for high-fidelity font synthesis without paired samples. Code and dataset are available at <span><span>https://github.com/nini739/FourCornerGAN</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":51011,"journal":{"name":"Digital Signal Processing","volume":"165 ","pages":"Article 105305"},"PeriodicalIF":2.9000,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Digital Signal Processing","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1051200425003276","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}

引用次数: 0

Abstract

Chinese character font generation poses unique challenges due to the complexity of glyph structures and the scarcity of paired training data. Existing methods for Chinese character font generation often suffer from issues like missing glyph formation and insufficient detail. To overcome these limitations, combining with the spatial glyph formation information, a novel encoding method based on the Four-Corner Number is proposed and integrated into CycleGAN to develop into FourCornerGAN to enhance structural representation in unpaired Chinese font generation, and a new Four-Corner Consistency Loss is introduced to guide both the generator and discriminator in preserving spatial glyph formation details. Extensive experiments demonstrate that FourCornerGAN significantly improves generation quality over baseline models, particularly in structural accuracy and visual consistency. This approach offers a promising solution for high-fidelity font synthesis without paired samples. Code and dataset are available at https://github.com/nini739/FourCornerGAN.

查看原文本刊更多论文

对非配对中文字体生成的字形形成增强

由于字形结构的复杂性和配对训练数据的稀缺性，汉字字体生成面临着独特的挑战。现有的汉字字体生成方法常常存在字形缺失、细节不足等问题。为了克服这些局限性，结合空间字形形成信息，提出了一种基于四角数的编码方法，并将其集成到CycleGAN中发展为FourCornerGAN，以增强非配对汉字字体生成中的结构表征，并引入了新的四角一致性损失来指导生成器和鉴别器保持空间字形形成细节。大量实验表明，与基线模型相比，FourCornerGAN显著提高了生成质量，特别是在结构精度和视觉一致性方面。这种方法为高保真字体合成提供了一种很有前途的解决方案，无需配对样本。代码和数据集可从https://github.com/nini739/FourCornerGAN获得。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Digital Signal Processing 工程技术-工程：电子与电气

CiteScore

5.30

自引率

17.20%

发文量

435

审稿时长

66 days

期刊介绍： Digital Signal Processing: A Review Journal is one of the oldest and most established journals in the field of signal processing yet it aims to be the most innovative. The Journal invites top quality research articles at the frontiers of research in all aspects of signal processing. Our objective is to provide a platform for the publication of ground-breaking research in signal processing with both academic and industrial appeal. The journal has a special emphasis on statistical signal processing methodology such as Bayesian signal processing, and encourages articles on emerging applications of signal processing such as: • big data• machine learning• internet of things• information security• systems biology and computational biology,• financial time series analysis,• autonomous vehicles,• quantum computing,• neuromorphic engineering,• human-computer interaction and intelligent user interfaces,• environmental signal processing,• geophysical signal processing including seismic signal processing,• chemioinformatics and bioinformatics,• audio, visual and performance arts,• disaster management and prevention,• renewable energy,