A natural representation of colors with textures

We present an automated method to transform an image to a set of binary textures that represent the intensities and colors of a full color original image. The foundation of our method is information preservation: creating a set of textures that allows for the reconstruction of the original image’s colors solely from the binarized representation. We present techniques to ensure that the textures created are not visually distracting, preserve the intensity profile of the images, and are natural in that they map sets of colors that are perceptually similar to patterns that are similar. The textures instantiated are fully reversible; the original images’ colors can be recreated from the single-bit per-pixel images. The approach uses deep-neural networks and is entirely self-supervised. The system yields aesthetically pleasing binary images when tested on a variety of image sources, including color and black and white photographs, clip art, and paintings.