Making Images from Images: Interleaving Denoising and Transformation

TL;DR

This paper introduces a novel method that interleaves denoising and transformation steps to generate new images by rearranging regions of existing images, enabling flexible and improved image transformations.

Contribution

It extends recent optical illusion generation techniques by learning image content and transformations simultaneously, allowing arbitrary source images to be transformed into new subjects.

Findings

Increasing regions improves results and simplifies the problem.

The method works in both pixel and latent spaces.

Extensions include using multiple source images and infinite copies.

Abstract

Simply by rearranging the regions of an image, we can create a new image of any subject matter. The definition of regions is user definable, ranging from regularly and irregularly-shaped blocks, concentric rings, or even individual pixels. Our method extends and improves recent work in the generation of optical illusions by simultaneously learning not only the content of the images, but also the parameterized transformations required to transform the desired images into each other. By learning the image transforms, we allow any source image to be pre-specified; any existing image (e.g. the Mona Lisa) can be transformed to a novel subject. We formulate this process as a constrained optimization problem and address it through interleaving the steps of image diffusion with an energy minimization step. Unlike previous methods, increasing the number of regions actually makes the problem easier and improves results. We demonstrate our approach in both pixel and latent spaces. Creative extensions, such as using infinite copies of the source image and employing multiple source images, are also given.