Research Of Image Compression Technology Based On Embedded Zero-tree Wavelet Coding

With the increasing popularity and quality of digital images, the data that need to be operated is more in a variety of storage and communications applications. Therefore, image compression is not only necessary to reduce the storage capacity and bandwidth requirements, but also need to be able to conduct compressed bit-stream extraction, editing and processing, in order to meet the special equipment and a wide range of applications, more flexible image compression technology is in an urgent need.Because of its multi-resolution, wavelet transform is widely used in image coding. It can facilitate the construction of an embedded bit stream to achieve an embedded coding. The Embedded Zero-tree Wavelet (EZW) based on Wavelet Transform Coding is currently recognized as a classic image compression coding method. This dissertation based on in-depth study of EZW coding method proposed an improved Embedded Zero-tree Wavelet coding method, and the simulation results showed the effectiveness of this method.This dissertation introduced the relevant knowledge and technology development process of digital image compression; according to different types of image compression the dissertation gived a variety of commonly used image compression coding method; the dissertation detailed description of the wavelet theory, by the wavelet transform for image, the dissertation analyze the characteristics of wavelet coefficients; the dissertation studied the nature of wavelet functions and its impact on the performance of image compression and selected the appropriate wavelet function of the image decomposition by the simulation results; the dissertation had in-depth study of the EZW coding method and showed the EZW coding process and the implementation process. According to the image after the wavelet transform sub-band high-frequency distribution and lack of classical EZW algorithm, a improved algorithm was proposed. The improved algorithm was to take on the most low-frequency sub-band predictive coding, and to use arithmetic coding after offsetting forecast errors; and set new threshold for High-frequency sub-band; according to the character of human visual system, high-frequency sub-band threshold was improved, according to the character of the main scanning table, that the frequency of a large sign directly with fewer binary code improved the efficiency of EZW coding. At the same time, this dissertation put forward to the EZW algorithm to improve in another way, the most low-frequency sub-band offset a fixed value, to reduce the difference of wavelet coefficients between low and high-frequency sub-band, and avoid the low-frequency sub-band need more coded bit-plane than high, and save the time of encoding and decoding.The dissertation proved that the improved algorithm in image compression performance and image compression quality was better than the original algorithm by experiment simulation.