In this paper we describe an algorithm to find an Iterated Function System with associated probabilities (IFS) in order to generate a given coloured image. The proposed optimization algorithm, based on the gradient method combined with Simulated Annealing, is much faster than the Simulated Annealing algorithm. Actually, more than one order if magnitude improvement in computational time has been gained. The overall algorithm can be viewed as a Simulated Quenching (SQ) method, which is quite efficient for the problem under consideration. Convergence curves are reported and discussed and an example of coded image is included.
Titolo: | Fractal encoding of digital images with an optimization algorithm |
Autori: | |
Data di pubblicazione: | 1994 |
Abstract: | In this paper we describe an algorithm to find an Iterated Function System with associated probabilities (IFS) in order to generate a given coloured image. The proposed optimization algorithm, based on the gradient method combined with Simulated Annealing, is much faster than the Simulated Annealing algorithm. Actually, more than one order if magnitude improvement in computational time has been gained. The overall algorithm can be viewed as a Simulated Quenching (SQ) method, which is quite efficient for the problem under consideration. Convergence curves are reported and discussed and an example of coded image is included. |
Handle: | http://hdl.handle.net/11368/2789124 |
ISBN: | 0889861900 |
Appare nelle tipologie: | 4.1 Contributo in Atti Convegno (Proceeding) |