Image analogies

Abstract

Figure 1 An image analogy. Our problem is to compute a new "analogous" image B that relates to B in "the same way" as A relates to A. Here, A, A , and B are inputs to our algorithm, and B is the output. The full-size images are shown in Figures 10 and 11. AbstractThis paper describes a new framework for processing images by example, called "image analogies." The framework involves two stages: a design phase, in which a pair of images, with one image purported to be a "filtered" version of the other, is presented as "training data"; and an application phase, in which the learned filter is applied to some new target image in order to create an "analogous" filtered result. Image analogies are based on a simple multiscale autoregression, inspired primarily by recent results in texture synthesis. By choosing different types of source image pairs as input, the framework supports a wide variety of "image filter" effects, including traditional image filters, such as blurring or embossing; improved texture synthesis, in which some textures are synthesized with higher quality than by previous approaches; super-resolution, in which a higher-resolution image is inferred from a low-resolution source; texture transfer, in which images are "texturized" with some arbitrary source texture; artistic filters, in which various drawing and painting styles are synthesized based on scanned real-world examples; and texture-by-numbers, in which realistic scenes, composed of a variety of textures, are created using a simple painting interface.Please see http://grail.cs.washington.edu/projects/image-analogies/ for additional information and results.While image analogies are clearly a desirable goal, it is not so clear how they might be achieved.

Perceptually inspired real-time artistic style transfer for video stream

“…Image analogy [10] is the first approach to stylize an input image using the texture of an example image. This approach synthesizes spatially coherent texture using multiscale autoregression.…”

Learning from Ideal Edge for Image Restoration

“…From the image analogy point of view, I bb is related to I b in the same way that I aa relates to I a . For better understanding, we adopt the notation of image analogy [9] although our algorithm differs substantially from that of Hertzmann in terms of both underlying rationale and implementation details. Let f a denotes the mapping from image I a to I aa (it can be represented by I a →I aa ), and f b denotes the mapping of I b →I bb , then " f a :: f b " represents that f a is identical to f b .…”

“…al [9] firstly searched the most matching point(p, q) in image I a for one point (i, j) in image I b , using the weighted distance over the feature vectors after the transformation of Gaussian Pyramid. Compared with Hertzmann's, the content of our training sample I a is more simplex.…”

“…In the hallucinating faces algorithm, the training images are restricted to a small subset containing similar face images. Hertzmann et al [9] proposed an algorithm using the concept of image analogy in the super-resolution restoration of color images. Nevertheless, both Baker's method and Hertzmann's method rely heavily on the selection of training samples, of which the applicability and reliability may be dubious.…”

Style Conversion of Cartoon Animation

“…Shiraishi et al [5] used the image moments of the color difference image to place rectangular brush stroke. Hertzmann et al [6] proposed an Image Analogy technique to process photo images into some artistic effects. More recently, DeCarlo and Santell [7] used a model of human perception and a record of a user's eye movements in looking at the photo to transform images into a line-drawing style.…”