StyleGAN2 Distillation for Feed-forward Image Manipulation

TL;DR

This paper introduces a method to distill StyleGAN2's image manipulation capabilities into a fast, feed-forward network, enabling real-time image editing with quality comparable to traditional optimization-based approaches.

Contribution

It presents a novel distillation technique that converts StyleGAN2 manipulations into an efficient image-to-image network, bypassing slow latent code optimization.

Findings

Generated images are comparable in quality to StyleGAN2 backpropagation.

The method enables real-time image manipulation tasks.

Applicable to various transformations like gender swap, aging, and style transfer.

Abstract

StyleGAN2 is a state-of-the-art network in generating realistic images. Besides, it was explicitly trained to have disentangled directions in latent space, which allows efficient image manipulation by varying latent factors. Editing existing images requires embedding a given image into the latent space of StyleGAN2. Latent code optimization via backpropagation is commonly used for qualitative embedding of real world images, although it is prohibitively slow for many applications. We propose a way to distill a particular image manipulation of StyleGAN2 into image-to-image network trained in paired way. The resulting pipeline is an alternative to existing GANs, trained on unpaired data. We provide results of human faces' transformation: gender swap, aging/rejuvenation, style transfer and image morphing. We show that the quality of generation using our method is comparable to StyleGAN2 backpropagation and current state-of-the-art methods in these particular tasks.