Painting Style-Aware Manga Colorization Based on Generative Adversarial Networks
Authors:
Yugo Shimizu,
Ryosuke Furuta,
Delong Ouyang,
Yukinobu Taniguchi,
Ryota Hinami,
Shonosuke Ishiwatari
Abstract:
Japanese comics (called manga) are traditionally created in monochrome format. In recent years, in addition to monochrome comics, full color comics, a more attractive medium, have appeared. Unfortunately, color comics require manual colorization, which incurs high labor costs. Although automatic colorization methods have been recently proposed, most of them are designed for illustrations, not for…
▽ More
Japanese comics (called manga) are traditionally created in monochrome format. In recent years, in addition to monochrome comics, full color comics, a more attractive medium, have appeared. Unfortunately, color comics require manual colorization, which incurs high labor costs. Although automatic colorization methods have been recently proposed, most of them are designed for illustrations, not for comics. Unlike illustrations, since comics are composed of many consecutive images, the painting style must be consistent. To realize consistent colorization, we propose here a semi-automatic colorization method based on generative adversarial networks (GAN); the method learns the painting style of a specific comic from small amount of training data. The proposed method takes a pair of a screen tone image and a flat colored image as input, and outputs a colorized image. Experiments show that the proposed method achieves better performance than the existing alternatives.
△ Less
Submitted 16 July, 2021;
originally announced July 2021.
