Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Missing Data Imputation GAN for Character Sprite Generation

Flávio Coutinho, Luiz Chaimowicz

TL;DR

This work presents a novel approach to character generation by framing the problem as a missing data imputation task, and proposes a generative adversarial networks model that receives the images of a character in all available domains and produces the image of the missing pose.

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.

A Missing Data Imputation GAN for Character Sprite Generation

TL;DR

This work presents a novel approach to character generation by framing the problem as a missing data imputation task, and proposes a generative adversarial networks model that receives the images of a character in all available domains and produces the image of the missing pose.

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.
Paper Structure (23 sections, 10 equations, 4 figures, 5 tables)

This paper contains 23 sections, 10 equations, 4 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Generative Adversarial Networks
  4. Image-to-Image Translation
  5. Related Work
  6. Sprite Generation
  7. Missing Data Imputation
  8. Architecture
  9. Objective Function
  10. Generator
  11. Discriminator
  12. Training Procedure
  13. Methodology
  14. Dataset
  15. Evaluation Metrics
  16. ...and 8 more sections

Figures (4)

  • Figure 1: Architecture of the proposed model. Left: The generator receives a character in the source domains and a label indicating the target, which is one-hot encoded, spatially spread, and concatenated with each input image. The inputs follow the encoder branches and are concatenated at the bottleneck layer, flowing into the unified decoder. Skip connections provide early outputs to the decoder. Right: The discriminator receives the image (real or fake) that must be distinguished and outputs $D_{adv}$ with the real/fake logit and $D_{dmn}$ with the probabilities of the image being part of each domain.
  • Figure 2: Comparison of input dropout (left) and replacement procedures (right) during training in the proposed model.
  • Figure 3: Sample images from the dataset showing different sizes/art styles (columns) facing four directions (rows).
  • Figure 4: Example images generated in different target domains. The columns show the source images, the target, the generation with the baselines using different source domains, and the generation using all sources with CollaGAN.