Application And Research Of Wavelet Analyse In Digital Image Compression

With the ever-growing multimedia technique, people are looking forward to new image compression techniques with better performance. Taking the storage and transfer of mass image data into consideration, not only could favorable image compression techniques achieve high compression ratio and fidelity ,they should also meet the demands of web applications, image compression need further detailed study, thus having developed into a specialized research field known as image coding. Frequently used major coding methods include predical code, transform code, vectol quantization, wavelet transform code, fractal code,ect.The wavelet transformation can provide a good mechanism to explore human visual system, while it removes the data correlation in image. So the image compression coding algorithm based on wavelet transformation has became researching hotspot in recent years. In this dissertation, a image compression coding method based on wavelet transformation and vector quantization is discussed in detail. Firstly, the general conception of image compression is presented, the brief developing history of image compression technique is reviewed and several image compression algorithm are introduced briefly. Then, the conception of wavelet transformation is introduced from the viewpoint of signal processing, and its application in image compression, including the selection of wavelet basis and the distribution of wavelet coeficients, is researched by the experimental ways. Then,we focus on Embedded Wavelet algorithm,include Embedded Zerotrees Wavelet algorithm,Set Partitioning in Hierarchical Trees algorithm and X-tree image coding algorithm. Through anatomizing EZW coding algorithm and taking the effects of human being's vision characteristics on quality of reconstituting image.A improved Zeretrees algorithm are put forward.At last a new image compression encoding algorithm based on wavelet transform and vector quantization is put forward. The experiment show, in very low bit rate, using the proposed algorithm, the reconstructed image is superior to that of EZW in both of perception and PSNR.