Hybrid Wavelet - Fractal Image Compression Method

with the development of multimedia technology, more high-performance image compression is expected to adven. As a new and excellent method used for signal analysis and processing, wavelet analysis is becoming one of the most important methods because of its perfect performance on both time and fuequency domain in recent years.Almost in the same time,the fractal compression is studied widely and maybe be used mostly on imgage compression in the future ,since a fractal block coder introducede by A.EJacquin is carried out by computer automatically.In recent years,fractal comprssion is applied to image decomposed by wavelet transform,because wavelet transform concentrate most energy in the approximate subband and there is some similarity between different resolution subband in the same direction. The hybrid algorithms of fractal and wavelet transformation coding have been developed and arte becoming popular.Some results of study have showed the efficiency and rationality of the hybrid method.This paper discusses it in such conditionFirst of all ,we discussed the definition and characteristics of wavelet transformation.,learned about existing image compression method based on the wavelet transformation .Then we implement the mostly used and most representative algorithems:EZW and SPIHT and compared the two algorithems.The followed part of this paper is about fractal theories and its use in image compression.The basic theories of fractal and the principles how it is applied to image compression is discussedAfter analyzing the shortcomings of the conventional fractal method ,a new improved method is put forward which is a fast fractal encoder based on the theory of nearest-neighbor search around the domain block whos variance is nearest to the variance of range block.The data of experiments showed the improved method saved a lot of time spent on coding and decoding and didn't reduce the qulity of image decoded.In the end of this paper, a hybrid wavelet _fractal (HWF)image coder is bringed forward.This coder combined the fractal coding and SPIHT coding to a decomposed image by wavelet transform.In the approximation subband,the improved fractal coding mentioned in the above paragraph is used;The SPIHT coding is used to the remained detail subband.As a consequence,fractal encoding time of original image is significantly reduced and the details,that would be lost without useing DWT ,are preserved in the 3k subbands.This detail information,encoded by modified SPIHT,is sent to the decoder,improving the reconstructed image quality.The result of experiment has improved the analysis above.