Japanese Kanji Font Style Transfer based on GAN with Unpaired Training

Sakai, Hiroki; Niino, Daisuke; Ijiri, Takashi

Japanese Kanji Font Style Transfer based on GAN with Unpaired Training

Files

005-006.pdf (465.43 KB)

supplementarymaterial.pdf (540.46 KB)

Date

2018

Authors

Sakai, Hiroki
Niino, Daisuke
Ijiri, Takashi

Publisher

The Eurographics Association

Abstract

To design a whole package of Japanese font is labor consuming, since it usually contains about 30k kanji characters. To support an efficient design process, this poster attempts to adopt a style transfer algorithm for font package completion. Given two font packages where one contains all characters and the other lacks a large part, we train CycleGAN to perform style transfer between the two packages and transfer the style from the former to the latter. To illustrate the feasibility of our technique, we performed style transfer experiments and achieved visually plausible results by using a relatively small training data set.

        @inproceedings{10.2312:pg.20181290
,
booktitle = {Pacific Graphics Posters
},
editor = {Fu, Hongbo and Ghosh, Abhijeet and Kopf, Johannes
},
title = {{Japanese Kanji Font Style Transfer based on GAN with Unpaired Training
}},
author = {Sakai, Hiroki and 
Niino, Daisuke and 
Ijiri, Takashi
},
year = {2018
},
publisher = {The Eurographics Association
},
ISBN = {978-3-03868-074-1
},
DOI = {10.2312/pg.20181290
}
}