Research And Simulation On Medical Image Enhancement Algorithm

Medical image is the important basis and source of modern medical diagnosis. Theveracity and promptness of medical diagnosis will improve in a great degree when medicalimaging technology is applied to the process of medical diagnosis. However, because of thelimitation of hardware and the influence of human tissue, the major of medical imagesdirectly from medical instruments have some problems such as blurring of detail, low contrast,heavy noise or artifact and so on. Besides, different medical images of different body partsthat are obtained from different medical instruments have different defects, for example, MRimages and CT images have heavy artifact, whereas ultrasound images have low contrast andblurry texture. And these factors aforementioned will lead to the blur of lesion andbackground in image, which will ulteriorly affect doctors’ orientation and diagnosis of lesion.So, it is necessary to utilize some enhancement algorithms to process medical image beforedoctor’s diagnosis, which will enhance the intelligibility of the local or designated spot ofimage to decrease misdiagnosis rate.The first part of this paper describes the development situation of medical imageenhancement at home and abroad. And then, section two introduces some medical imageenhancement algorithms and analyzes the defects and problems of these commonly usedalgorithms when we process image in reality. Next, the thesis proposed some improvedmedical image enhancement methods on the basis of these analysis, what’s more, thesubsequent section use some medical images to test the effectiveness of these proposedmethods. The main work of this paper includes the following parts:1. It studies an improved enhancement method on the basis of wavelet transform toadaptively enhance the contrast of X-ray medical image with low contrast. Afterwards, wavelet basis will be chosen by using some special condition, and then, a large number ofwavelet coefficients with rich image information will be obtained after the paper processescoefficients of every direction that are produced by wavelet transform by employing certainadaptive filtering. At last, we apply the inverse wavelet transform to map these modifiedcoefficients to the space domain to get medical image with proper contrast, which means theresult image will provide more help for doctors’diagnosis.2. Further research of medical image un-sharp masking enhancement is made. This papertakes the human visual characteristic into consideration, that the human visual system is moresensitive to intensity in the mid intensity range than in the dark or bright range to avoid overenhancement. In the meantime, the local difference curvature of image can be used todistinguish edges and details region from flat region so that we can get the result medicalimages that have rich details, higher signal to noise ratio, which provides more conveniencefor subsequent analysis and processing.