Research On Medical Image Processing Based On Wavelet Transform

In modern medical science, medical image technology has already become doctor's essential means and tool. The quality of medical image influences directly doctor's diagnosis and treating to the disease. The medical images often are mixed with various kinds of noises during the process of getting. Such as CT image, when the attenuation coefficient is no appreciable distinction between pathological organization and normal organization, the focus will be not distinguished by high noise CT. CT diagnosis can draw a conclusion based on the normal or abnormal information provided by CT image. The accordant degree of diagnosis lies on imagequality. High quality CT image is the precondition of accurate diagnosis. So it is necessory to reduce the influence of the noise as much as possible.Wavelet transform has been developed in recent decade, Using wavelet transform to processing medical image has already become a great researche and developement focus in medical process. Wavelet transform is widely applied to the field of medical image. Median filtering is a nonlinear denoising algorithm that can better restrict pulse noise and better keep image edge.This paper studies mainly CT image denoising processing. CT images included Gauss noise, pulse noise and mix noise of both are denoised by traditional denoising algorithm, such as median filtering (space domain algorithm), Butterworth lowpass filter and threshold method of wavelet transform. Then traditional denoising algorithms are improved. There are six improved algorithms, such as combination of space domain algorithm and frequency domain algorithm, combination of median filtering and Butterworth lowpass filter and combination threshold method of wavelet transform. That is to say, algorithms include first median filteringand Butterworth lowpass filter later, first Butterworth lowpass filter and median filtering later, first Butterworth lowpass filter and threshold method of wavelet transform later, first threshold values of wavelet transform and Butterworth lowpass filter later, first threshold method of wavelet transform and median filtering and threshold medthod of wavelet transform later and first median filtering and threshold medthod of wavelet transform later. When CT image include Gauss noise, pulse noise and both of them are denoised separately. It is proved that the method of first median filtering and others later is better than others. Among improved six algorithms, the algorithm of first median filtering and threshold mothod of wavelet transform later is best. Following the result of experiment, this paper presents a new denoising algorithm (new algorithm), which uses the principle of space field algorithm in frequency domain algorithm. Firstly, noised CT image is treated using median filtering. Secondly, filtered image is decomposed by wavelet transform and product wavelet coefficient matrix. Employing the theory of median filtering to deal with wavelet coefficient matrix, a new wavelet coefficient matrix can be prduced.The image can be reconstructed by using the new wavelet coefficient matrix. Finally the image is denoised by wavelet transform threshold algorithm and generate, a new denoised image. Compared the algorithm with others, the visual effect of image is best and MSB is least. It is a ideal algorithm that CT images including Gauss noise, pulse noise and both of them are denoised.With the rapid development of computer technology and modern communications technology, remote diagnose and cure become possible. But the medical image is large in data amount, the memory space is big and it is slow to transmit. Thus it is needed to compress the image. This paper utilizes the wavelet transform to research the compression of the medical image. Colored or black-and-white images are compressed by employing the wavelet decomposing algorithm and the global threshold algorithm separately. The experimental result indicates the global threshold results are better than the result of wavelet packets algorithm.The technology of image fusion not only combines various kinds of medical image information organically, but also opens up new thinking for modern medical clinical diagnosis at the sametime and offers the new standard. It is not only a research focus in recent years, but also a front subject of the contemporary medical image process field. This paper utilizes the wavelet transform to discuss preliminary to the medical image fusion.