Auto-painter: Cartoon Image Generation from Sketch by Using Conditional Generative Adversarial Networks

TL;DR

This paper introduces Auto-painter, a cGAN-based model that automatically generates colorful cartoon images from sketches, allowing user preferences and outperforming existing methods in sketch-to-image synthesis.

Contribution

The paper presents a novel cGAN model for sketch-to-color image generation that incorporates user preferences and demonstrates superior performance over existing techniques.

Findings

Auto-painter effectively colors sketches with proper and user-indicated colors.

The model outperforms existing image-to-image translation methods.

Experimental results validate the model's capability in sketch-based cartoon image generation.

Abstract

Recently, realistic image generation using deep neural networks has become a hot topic in machine learning and computer vision. Images can be generated at the pixel level by learning from a large collection of images. Learning to generate colorful cartoon images from black-and-white sketches is not only an interesting research problem, but also a potential application in digital entertainment. In this paper, we investigate the sketch-to-image synthesis problem by using conditional generative adversarial networks (cGAN). We propose the auto-painter model which can automatically generate compatible colors for a sketch. The new model is not only capable of painting hand-draw sketch with proper colors, but also allowing users to indicate preferred colors. Experimental results on two sketch datasets show that the auto-painter performs better that existing image-to-image methods.