Fractal-based methods in imaging Original Research Article

Through a procedure called the LIFSM algorithm, based on the Collage Theorem, it is possible to represent an image by a set of data, known as a fractal code, from which we may construct an approximation of the image. This algorithm is largely based upon the idea of matching larger portions of an image, called “parent blocks,” to smaller portions called “child blocks” in order to minimize a quantity known as the “collage distance.” We modify the LIFSM algorithm through the consideration of “near-optimal pairings.” We develop the theory, discuss applications and present examples in image denoising, image reconstruction, and edge detection.