A Missing Data Imputation GAN for Character Sprite Generation

TL;DR

This paper introduces a GAN-based method for imputing missing character sprite images in pixel art, automating pose generation to assist artists and improve sprite creation efficiency.

Contribution

It presents a novel GAN framework for missing data imputation in pixel art sprite generation, handling multiple missing images and enhancing existing architectures.

Findings

Achieved comparable or better results than state-of-the-art with multiple missing images.

Demonstrated effectiveness in generating missing character poses.

Evaluated architectural improvements for better imputation quality.

Abstract

Creating and updating pixel art character sprites with many frames spanning different animations and poses takes time and can quickly become repetitive. However, that can be partially automated to allow artists to focus on more creative tasks. In this work, we concentrate on creating pixel art character sprites in a target pose from images of them facing other three directions. We present a novel approach to character generation by framing the problem as a missing data imputation task. Our proposed generative adversarial networks model receives the images of a character in all available domains and produces the image of the missing pose. We evaluated our approach in the scenarios with one, two, and three missing images, achieving similar or better results to the state-of-the-art when more images are available. We also evaluate the impact of the proposed changes to the base architecture.