In-Domain GAN Inversion for Faithful Reconstruction and Editability

TL;DR

This paper introduces in-domain GAN inversion, combining a domain-guided encoder and regularized optimization to improve image reconstruction and semantic editing without retraining, revealing a trade-off between quality and editability.

Contribution

It proposes a novel in-domain GAN inversion method that enhances image reconstruction and editing capabilities by regularizing the latent space of pre-trained GANs.

Findings

The encoder structure significantly impacts inversion quality.

The starting point of inversion affects the reconstruction and editability.

There is a trade-off between reconstruction fidelity and semantic editability.

Abstract

Generative Adversarial Networks (GANs) have significantly advanced image synthesis through mapping randomly sampled latent codes to high-fidelity synthesized images. However, applying well-trained GANs to real image editing remains challenging. A common solution is to find an approximate latent code that can adequately recover the input image to edit, which is also known as GAN inversion. To invert a GAN model, prior works typically focus on reconstructing the target image at the pixel level, yet few studies are conducted on whether the inverted result can well support manipulation at the semantic level. This work fills in this gap by proposing in-domain GAN inversion, which consists of a domain-guided encoder and a domain-regularized optimizer, to regularize the inverted code in the native latent space of the pre-trained GAN model. In this way, we manage to sufficiently reuse the knowledge learned by GANs for image reconstruction, facilitating a wide range of editing applications without any retraining. We further make comprehensive analyses on the effects of the encoder structure, the starting inversion point, as well as the inversion parameter space, and observe the trade-off between the reconstruction quality and the editing property. Such a trade-off sheds light on how a GAN model represents an image with various semantics encoded in the learned latent distribution. Code, models, and demo are available at the project page: https://genforce.github.io/idinvert/.