Research On Improved Embedded Zerotree Coding Algorithm Based On Wavelet Transform

Image is the main carrier of information transmission in modern society. Because of the limitation in transmission bandwidth and storage capacity, image compression becomes one of the essential techniques in the information transmission system. The wavelet transformation technology based on its good spatial frequency partial characteristic and the eye visual characteristic, has obtained widespread application and research in the image code domain.A deep research on static image compression is presented in this paper. First, elaborated the rationale of compression code algorithm based on the wavelet transform; then a thorough research about embedded zero-tree wavelet algorithm is elaborated, introduced the theory basis of this method based on zero tree structure, showed the realization process of this method, and analyzed the good and bad points based on the EZW algorithm; finally an improvement algorithm is proposed in view of the insufficiency of EZW algorithm. The improvement algorithm mainly includes following several aspects: adopting lifting wavelet transform, realized lossless and reversible transformation in the true significance; using the DPCM method to code the lower frequency individually to reduce the distortion rate; setting different quantization threshold in different direction, enlarging the quantification threshold value gradually from the lower resolution to the higher resolution, and adopting the quantization threshold based on the integers'square; using the method which the zero-tree coding and entropy coding unifies, enhanced the efficiency effectively. At last, a satisfying result has obtained in the experiment about test picture Lena and woman, and carries on the comparison with the SPIHT and SPECK, thus confirmed the improvement algorithm accuracy and feasibility.