Recovering fractal model by optimization using orthogonal tables and fixed points

This paper addresses the problem of recovering the fractal model and its parameters from an object or an image. We propose a method to solve this problem by use of the iterated functions system (IFS) and optimization with help of the orthogonal tables, which are established from the experimental plans of Taguchi. This method allows representing objects of different shapes by fractal parameter coding and could thus be used for image compression. As is well known, in general a fractal model contains many parameters and the solution found by optimization depends much on the initial solution proposed, which is one of the most difficult problems in fractal modeling. Our approach gives a solution to this problem by using the fixed points and the new algorithm presented.