A Novel Medical Image Enhancement Algorithm Based On Ridgelet Transform

It is no longer difficult to access various types of medical images with the development of science and technology,especially the development of medical technology and imaging technology.Medical images play an important role in the diagnosis of modern diseases,they can help clinicians to diagnose and analyze the disease.However,the quality of medical images will be affected due to the impact of data acquisition equipment,noise and other factors.In order to assist clinicians on their work much better and provide more clearly and accurately medical images,medical image enhancement processing has been becoming an important problem in the filed of medical image processing.This paper focuses on the methods of medical image enhancement,and a novel enhancement algorithm of medical images based on ridgelet transform is proposed combining the theory of sparse representation and multi-wavelet transform.The main works of this thesis are as follows.Firstly,the methods of medical image enhancement in the spatial domain were studied.These methods are the direct method to process the gray level of the original image,mainly including gray level transformation,histogram equalization and spatial filtering.Secondly,the theory of wavelet transform and ridgelet transform were studied.Then,a novel medical image enhancement algorithm based on ridgelet transform is proposed.The main processes of the proposed algorithm are as follows: first of all,the medical image is processed by Radon transform,then the result of Radon transform is transformed by one dimension multi-wavelet transform.After this,the coefficients of the multi-wavelet transform are enhanced by enhancement factor.On the basis of this,the medical image is reconstructed by inverse multi-wavelet transform and inverse Radon transform.Finally,the reconstructed image is processed by gray level transformation to get much better enhanced image.Experimental results show that the proposed method not only can enhance an image's details but also hold the image's edge features effectively.