Research Of Medical Image Enhancement Based On Block Finite Ridgelet Transform

In recent years, with the development of medical technology, computer technology andsensor technology, it is possible to obtain all kinds of medical images, which is very importantto assist clinicians in the diagnosis and analysis of disease. However, due to the impact of dataacquisition equipment, lighting conditions, noise and other factors, medical images tend to bedegraded in quality. The factors influencing the quality of medical images are mainly on lowcontrast and high level of noise. In order to provide more clearly, accurately and desirablymedical images for clinician, combined with image sparse representation theory, this thesisfocuses on the research of medical image enhancement processing theory and its applicationtechnology, and then a medical image enhancement algorithm based on block ridgelettransform is proposed in this thesis.Firstly, this thesis researches on the methods of spatial domain image enhancement, andthen discussed the enhancement methods of gray-scale transformation, histogram equalization,spatial filtering etc. In many cases, medical images need to enhance the detail information.Therefore, only by using the spatial domain enhancement methods are often insufficient toachieve the desired effect of enhancement. Secondly, based on the wavelet theory, this thesisdiscusses one-dimensional signal and two-dimensional signal’s wavelet decomposition.Thirdly, this thesis studies on one of the image sparse representation theories called ridgelet.Since ridgelet has significantly superior for dealing with the signals which enjoy theproperties of directional singular feature, and then applies it in the field of image enhancement.Finally, a medical image enhancement algorithm based on block finite ridgelet transform isproposed in this thesis. In the algorithm, a piecewise nonlinear function, a modular squarethreshold function and wiener filtering method are used in each sub block of ridgelettransform domain coefficients. Then using method of gray-scale transformation for thereconstructed medical image in order to increase the image’s contrast. Some numericalexperiment results show that the proposed algorithm can obtain good enhancement effect formedical images.