Application Of Wavelet Transform In Image Coding

In recent twenty years,the wavelet analysis and wavelet transform develop very fast. They have been used into many fields in technique.Becuse of their good local performance in both time(space) and frequency, they are particular for the analysis of nonstationary signals. In the area of image, waveles have been used very well.The article is mainly about image compress using wavelets.First,we give the information theory that is needed in image comperssion, and introduce the knowledge of image compression.In the article,we simply introduce the wavelet transform, analyze the two-dimension wavelet decomposition and construct algorithm, focus on introducing the basic theory of two-demension multi-resolution.In addition, we give the feature of wavelet transform.In the article,we introduce the lifting wave transform importantly,and give a new method using the algebra coefficients equation to solve the lifting wavelet coeffienrs, this method is more concise than the Euclid method, then we have given the method of deriving the coefficients of lifting scheme of the Daubiechies biorthogonal wavelet. Compared with Mallat algorithm, the lifting algorithm is more superiority.At last, we divide the image into many sub-image using the feature of multi-resolution, according to the feature of the energe, concentrated in the lowest frequncy subband, the image is coded by segment to gray scale. For other subbands, we use a new method based on embedded Zero-trees wavlet, we denoise and delete the coefficients of the noise. And add two special zerotrees PZTR and NZTR.In addition, we use the quantify factors to quantify the lowest four subbands, then we lower the quantity of isolated zero. According the experiment, the method can improve the efficience:lower the time, higer the PSNR.of image coding in the same compression ratio.