Navigating the GAN Parameter Space for Semantic Image Editing

TL;DR

This paper explores the generator parameter space of GANs, particularly StyleGAN2, revealing interpretable directions that enable novel semantic image editing beyond traditional latent space manipulations.

Contribution

It introduces methods to discover interpretable directions in GAN generator parameters, expanding the range of visual effects for image editing beyond latent code transformations.

Findings

Identified interpretable directions in generator parameters for semantic editing.

Demonstrated editing capabilities on both synthetic and real images.

Provided code and models for community use.

Abstract

Generative Adversarial Networks (GANs) are currently an indispensable tool for visual editing, being a standard component of image-to-image translation and image restoration pipelines. Furthermore, GANs are especially useful for controllable generation since their latent spaces contain a wide range of interpretable directions, well suited for semantic editing operations. By gradually changing latent codes along these directions, one can produce impressive visual effects, unattainable without GANs. In this paper, we significantly expand the range of visual effects achievable with the state-of-the-art models, like StyleGAN2. In contrast to existing works, which mostly operate by latent codes, we discover interpretable directions in the space of the generator parameters. By several simple methods, we explore this space and demonstrate that it also contains a plethora of interpretable directions, which are an excellent source of non-trivial semantic manipulations. The discovered manipulations cannot be achieved by transforming the latent codes and can be used to edit both synthetic and real images. We release our code and models and hope they will serve as a handy tool for further efforts on GAN-based image editing.