Image restoration based on estimation of fractal structure

In this paper, image restoration under unknown degradation process (hard to make a mathematical model of the degradation process) will be treated, and a novel approach adopting fractal coding is proposed. The method begins, as a pre-processing, with the construction of a table of range-domain correspondences on a degraded image using known pairs of an original image and its degraded image. In the restoration phase, the fractal structure of an unknown image is, first, estimated from an input image by using the table, and next, a restored image is generated by the decoding process. The fractal decoding procedure is evolved by the concept of the fuzzy vector quantization to diminish the distortion of a decoded image. Features and effectiveness are examined through some simulations