Method for image data reduction by fractal image encoding with encoder and decoder for carrying out said method

摘要

The method involves reducing luminance and/or colour data of a digital image signal by means of fractal image coding, at which each image is subdivided into range blocks with respectively n*n image points. A domain block is searched for each range block, which is mappable onto the range block under application of a transformation function with most little possible deviation. Basic geometric functions are searched both in the range and domain blocks, which approximate the data structure of the respective block as good as possible. A difference value is derived from the function to describe a residual roughness of the block. The search for the most appropriate domain block is carried out based on those difference values.