Abstract

Deep generative models have become increasingly effective at producing realistic images from randomly sampled seeds, but using such models for controllable manipulation of existing images remains challenging. We propose the Swapping Autoencoder, a deep model designed specifically for image manipulation, rather than random sampling. The key idea is to encode an image with two independent components and enforce that any swapped combination maps to a realistic image. In particular, we encourage the components to represent structure and texture, by enforcing one component to encode co-occurrent patch statistics across different parts of an image. As our method is trained with an encoder, finding the latent codes for a new input image becomes trivial, rather than cumbersome. As a result, it can be used to manipulate real input images in various ways, including texture swapping, local and global editing, and latent code vector arithmetic. Experiments on multiple datasets show that our model produces better results and is substantially more efficient compared to recent generative models.

Landscape Mixer in Photoshop

Swapping Autoencoder inspired the development of the Landscape Mixer feature in Photoshop, released Oct 26th, 2021. It was introduced at the Keynote Speech of Adobe MAX 2021.

3-Minute Video Intro of the NeurIPS 2020 Paper

Paper

T. Park, J.Y. Zhu, O. Wang, J. Lu,
E. Shechtman, A. A. Efros, R. Zhang.
Swapping Autoencoder for
Deep Image Manipulation.
NeurIPS, 2020.

[Paper] | [Bibtex]

Acknowledgements

We thank Nicholas Kolkin for the helpful discussion on the automated content and style evaluation. We thank Jeongo Seo and Yoseob Kim for advice on the user interface. We thank Tongzhou Wang, William (Bill) Peebles, and Yu Sun for the discussion about disentanglement. Taesung Park is supported by a Samsung Scholarship and an Adobe Research Fellowship, and much of this work was done as an Adobe Research intern. This research was supported in part by an Adobe gift.