This work contains a general description of fractal image compression, an introduction to fractals and shows a new projection of acceleration in fractal image decoding. A new method based on modification of start conditions by decoding process was presented. The convergence speed of decoding is faster then convergence speed of standard method. The efficiency of this approach was verified on static gray-scale image LENNA.256 and on static color image LENNA.256 too. Three-components RGB model was used in verification of color image. Fractal image coding is a new and modern technique for lossy image compression. Some images can be compressed over 80:1 or higher. For a higher compress ratio the application of this method is very effective. Fractal Compression is an asymmetric process, compre-ssion taking a long time compared with decompression. There are several variations of fractal image compression for example fast coding methods, modifications for higher compression ratio, hybrid methods which used properties of vector quantization, image transforms, image prediction, etc. It is possible to modify decoding algorithm too. The modifications can be realized for an acceleration of standard algorithm, for acceleration of convergence to attractor or for elimination of iterative decoding process.
This diploma work describes one possible approach to acceleration of decoding algorithm. It is based on modification of start conditions by decoding of gray-scale and color images.