Study On The Detection In Medical Lesions

It is effective in helping doctors make the diagnosis of lesions to segment and extract the region of interest from medical images. Combining previous methods of extracting Region of Interest(ROI) in the ultrasound and X-ray images and considering the ultrasound gallstone image and mammography, here, it proposed a new ultrasound gallstones segmentation method,which employed the variational level set method. Moreover, it presents an improved level set model to detected lesion of mammography, automatically and accurately. In the ultrasound images of gallstones, the ROI are gallstones. The ROI are masses in mammography. According these two medical imaging, this paper mainly focuses on the following aspects:1. The principle of level set algorithm was shown, advantages of level set method also with disadvantages were comprehensively described, currently.2. A variational level set method had been demonstrated, and applying it to gallstone extraction in ultrasound imaging. Compared to PCNN model, the experimental results are presented to show that the variational level set method has gallstone detection results.3. A new breast mass segmentation process is proposed including preprocessing, initialization and mass extraction. As well known, it is difficult to robustly achieve mammogram images segmentation due to low contrast between normal and lesion tissues or the noise in such images. Therefore, PCNN algorithm is employed to detect the initial contours. And the extracted contours as the initial zero level set contours, the improved level set evolution is performed to segment the mammograms to get the final result. The new proposed algorithm improves external energy of level set method in terms of low contrast images. Besides, little work has been done on the initial zero level set contours based on PCNN algorithm for latterly level set evolution by now. In order to present that our proposed algorithm is more excellent, we make comparison experiments by using Li’s and C-V model. The experimental results demonstrate that our proposed approach can potentially obtain better masses detection results in terms of accuracy and specificity. Ultimately, this algorithm could lead to increase both sensitivity and specificity of the physicians’ interpretation of mammograms in clinical practice.