Research On Medical Image Compression Based On Wavelet Transform

With the development and the application of information technology, digital constructions of medical institutes become maturely; more are more digits medical imaging equipments are being used, and medical image in diagnosing process plays more and more important role. The contradiction between huge medical image data,storage space and transmit width makes the compression of medical image become necessity. Consequently, medical image compression researching is an important part of digits image processing.As one kind of the Grey images with which Medical image has the same characters, however, Medical images has it particularity, such as it should meets the requirement of clinic diagnosis. According to the characters above, this paper aims to improve the quality of reconstructed image by lifting the efficient of coding algorithm, with the precondition of guaranteeing the image compression ratio.Firstly, based on Wavelet transform theory, this paper makes deeply analyze about some mainly issues that produced in the process of Wavelet transforming. By Wavelet bases principle in Image compression, and by experiment data comparison of Standard grey image and Medical image, we selected the most effective Wavelet bases. Then use this selected Wavelet bases to compare data from different layer of the medical image. Lastly, select the better method according to the character of selected Wavelet bases and some kinds of Border extension methods.Secondly, this paper provides an improved SPIHT algorithm to handle some flaw in original edition. Image that after be Wavelet transferred, focus almost parts of its energy in low-frequency sub region in which located relatively high factor Wavelet, and this leads to many zero trees produced at the beginning of scanning, and results in outputting some unnecessary symbols placeholders. Meanwhile, it exits redundancy on divide set judgment of SPIHT algorithm. This paper improved the coding efficient by changing the zero construction rules and the divide set judgment rule of original algorithm, so it can decrease the number of zero tress produced at the beginning of scanning, and also decrease the redundancy produced by divide set judgment. At last, this paper implements an image compression coding system by MATLAB, and emulates the EZW,SPIHT and improved algorithm on standard images and different kinds of medical images. Experiments on standard images and different medical images show that the PSNR of improved algorithm is better than that of traditional zero tree coding methods at the same ratio.