GlyphGAN: Style-consistent font generation based on generative adversarial networks

Research output: Contribution to journalArticle

Abstract

In this paper, we propose GlyphGAN: style-consistent font generation based on generative adversarial networks (GANs). GANs are a framework for learning a generative model using a system of two neural networks competing with each other. One network generates synthetic images from random input vectors, and the other discriminates between synthetic and real images. The motivation of this study is to create new fonts using the GAN framework while maintaining style consistency over all characters. In GlyphGAN, the input vector for the generator network consists of two vectors: character class vector and style vector. The former is a one-hot vector and is associated with the character class of each sample image during training. The latter is a uniform random vector without supervised information. In this way, GlyphGAN can generate an infinite variety of fonts with the character and style independently controlled. Experimental results showed that fonts generated by GlyphGAN have style consistency and diversity different from the training images without losing their legibility.

Original languageEnglish
Article number104927
JournalKnowledge-Based Systems
Volume186
DOIs
Publication statusPublished - Dec 15 2019

Fingerprint

Neural networks
Generator

All Science Journal Classification (ASJC) codes

  • Management Information Systems
  • Software
  • Information Systems and Management
  • Artificial Intelligence

Cite this

GlyphGAN : Style-consistent font generation based on generative adversarial networks. / Hayashi, Hideaki; Abe, Kohtaro; Uchida, Seiichi.

In: Knowledge-Based Systems, Vol. 186, 104927, 15.12.2019.

Research output: Contribution to journalArticle

@article{7929ceee6c7545aea43c7f55cc450554,
title = "GlyphGAN: Style-consistent font generation based on generative adversarial networks",
abstract = "In this paper, we propose GlyphGAN: style-consistent font generation based on generative adversarial networks (GANs). GANs are a framework for learning a generative model using a system of two neural networks competing with each other. One network generates synthetic images from random input vectors, and the other discriminates between synthetic and real images. The motivation of this study is to create new fonts using the GAN framework while maintaining style consistency over all characters. In GlyphGAN, the input vector for the generator network consists of two vectors: character class vector and style vector. The former is a one-hot vector and is associated with the character class of each sample image during training. The latter is a uniform random vector without supervised information. In this way, GlyphGAN can generate an infinite variety of fonts with the character and style independently controlled. Experimental results showed that fonts generated by GlyphGAN have style consistency and diversity different from the training images without losing their legibility.",
author = "Hideaki Hayashi and Kohtaro Abe and Seiichi Uchida",
year = "2019",
month = "12",
day = "15",
doi = "10.1016/j.knosys.2019.104927",
language = "English",
volume = "186",
journal = "Knowledge-Based Systems",
issn = "0950-7051",
publisher = "Elsevier",

}

TY - JOUR

T1 - GlyphGAN

T2 - Style-consistent font generation based on generative adversarial networks

AU - Hayashi, Hideaki

AU - Abe, Kohtaro

AU - Uchida, Seiichi

PY - 2019/12/15

Y1 - 2019/12/15

N2 - In this paper, we propose GlyphGAN: style-consistent font generation based on generative adversarial networks (GANs). GANs are a framework for learning a generative model using a system of two neural networks competing with each other. One network generates synthetic images from random input vectors, and the other discriminates between synthetic and real images. The motivation of this study is to create new fonts using the GAN framework while maintaining style consistency over all characters. In GlyphGAN, the input vector for the generator network consists of two vectors: character class vector and style vector. The former is a one-hot vector and is associated with the character class of each sample image during training. The latter is a uniform random vector without supervised information. In this way, GlyphGAN can generate an infinite variety of fonts with the character and style independently controlled. Experimental results showed that fonts generated by GlyphGAN have style consistency and diversity different from the training images without losing their legibility.

AB - In this paper, we propose GlyphGAN: style-consistent font generation based on generative adversarial networks (GANs). GANs are a framework for learning a generative model using a system of two neural networks competing with each other. One network generates synthetic images from random input vectors, and the other discriminates between synthetic and real images. The motivation of this study is to create new fonts using the GAN framework while maintaining style consistency over all characters. In GlyphGAN, the input vector for the generator network consists of two vectors: character class vector and style vector. The former is a one-hot vector and is associated with the character class of each sample image during training. The latter is a uniform random vector without supervised information. In this way, GlyphGAN can generate an infinite variety of fonts with the character and style independently controlled. Experimental results showed that fonts generated by GlyphGAN have style consistency and diversity different from the training images without losing their legibility.

UR - http://www.scopus.com/inward/record.url?scp=85070695727&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070695727&partnerID=8YFLogxK

U2 - 10.1016/j.knosys.2019.104927

DO - 10.1016/j.knosys.2019.104927

M3 - Article

AN - SCOPUS:85070695727

VL - 186

JO - Knowledge-Based Systems

JF - Knowledge-Based Systems

SN - 0950-7051

M1 - 104927

ER -