GANravel: User-Driven Direction Disentanglement in Generative Adversarial Networks

TL;DR

GANravel is a user-driven tool that enhances control over GAN editing directions, enabling users to iteratively improve disentanglement and produce high-quality creative outputs, outperforming existing methods.

Contribution

It introduces a novel user-driven disentanglement tool that complements existing GAN architectures, allowing iterative improvement and better control over editing directions.

Findings

Users effectively disentangled directions with GANravel.

GANravel outperformed state-of-the-art baselines in disentanglement.

High-quality creative outputs like memes were produced using GANravel.

Abstract

Generative adversarial networks (GANs) have many application areas including image editing, domain translation, missing data imputation, and support for creative work. However, GANs are considered 'black boxes'. Specifically, the end-users have little control over how to improve editing directions through disentanglement. Prior work focused on new GAN architectures to disentangle editing directions. Alternatively, we propose GANravel a user-driven direction disentanglement tool that complements the existing GAN architectures and allows users to improve editing directions iteratively. In two user studies with 16 participants each, GANravel users were able to disentangle directions and outperformed the state-of-the-art direction discovery baselines in disentanglement performance. In the second user study, GANravel was used in a creative task of creating dog memes and was able to create high-quality edited images and GIFs.