StyleGAN of All Trades: Image Manipulation with Only Pretrained StyleGAN

TL;DR

This paper demonstrates that a pretrained StyleGAN can be used for diverse image manipulation tasks without additional training or architecture, achieving competitive results across multiple applications.

Contribution

It reveals the spatial properties of StyleGAN enabling various tasks with simple operations, eliminating the need for task-specific models or training.

Findings

Comparable performance to state-of-the-art methods on multiple tasks

Effective manipulation using only pretrained StyleGAN and simple operations

Applicable to any existing pretrained StyleGAN model

Abstract

Recently, StyleGAN has enabled various image manipulation and editing tasks thanks to the high-quality generation and the disentangled latent space. However, additional architectures or task-specific training paradigms are usually required for different tasks. In this work, we take a deeper look at the spatial properties of StyleGAN. We show that with a pretrained StyleGAN along with some operations, without any additional architecture, we can perform comparably to the state-of-the-art methods on various tasks, including image blending, panorama generation, generation from a single image, controllable and local multimodal image to image translation, and attributes transfer. The proposed method is simple, effective, efficient, and applicable to any existing pretrained StyleGAN model.