Swapping Autoencoder for Deep Image Manipulation
This addresses the problem of precise image editing for applications like graphics and design, offering a more efficient and effective solution than existing generative approaches.
The paper tackles the challenge of controllable image manipulation with deep generative models by proposing the Swapping Autoencoder, which encodes images into independent structure and texture components to enable realistic swaps, resulting in better performance and substantially higher efficiency compared to recent models.
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.