GANime: Generating Anime and Manga Character Drawings from Sketches with Deep Learning

TL;DR

This paper evaluates various deep learning models for converting sketches into detailed, colorized anime and manga character images, identifying C-GAN as the most effective approach for high-quality, high-resolution results.

Contribution

It compares multiple image-to-image translation models and demonstrates that C-GAN outperforms others in generating realistic anime character images from sketches.

Findings

C-GAN produces the highest quality images.

C-GAN achieves high-resolution outputs.

Model assessment is both qualitative and quantitative.

Abstract

The process of generating fully colorized drawings from sketches is a large, usually costly bottleneck in the manga and anime industry. In this study, we examine multiple models for image-to-image translation between anime characters and their sketches, including Neural Style Transfer, C-GAN, and CycleGAN. By assessing them qualitatively and quantitatively, we find that C-GAN is the most effective model that is able to produce high-quality and high-resolution images close to those created by humans.