Research On Application Of Wavelet Transform In Medical Image Compression

With the fast development and widely used of science and technology, digital image storage and transmission played an important role in the society. In medical fields, medical imaging equipments were applied wildly in medical systems, and produced a great deal medical image data, and this put forward higher requirement to medical image storage and transmission. Therefore, the research of medical image compression coding is an important subject in digital image processing.The techniques of image coding based on wavelet transform are newly developed coding methods, they have not only the advantages of traditional coding techniques, but also good localization in frequency and time domains and multi resolution characteristic, so they have a rapid development and a wide application. This paper researches wavelet transform theory, analyzes the methods of wavelet construction, some of the properties of wavelet bases, the determination of wavelet transform scales and some general methods for border processing in wavelet transform.The traditional zero tree coding methods based on wavelet transform, such as EZW, SPIHT, have some disadvantages. Analyzing SPIHT algorithms in detail, during initialization, the number of trees it constructed is large. In the several scans at the beginning, most of the trees are zero trees, as the threshold is big. It outputs some unnecessary sign bits. To compensate for the shortcomings, a new coding method is adopted, to improve the traditional zero tree coding methods. In order to combine the trees constructed during initialization, we introduce a new type of tree to hold as many coefficients as possible, this decreases the number of trees.Experiments on standard images and different medical images show that the PSNR of improved algorithm is better than that of original SPIHT at the same ratio.