| With the development of medical imaging technogy,Image processing in medical research and clinical application are more widely.Medical images usually consist of the region of interest and background,the regions of interest contain important diagnostic information,although it may be small compared with the entire image,but the cost of its improper description is very high.The purpose of image segmentation is extract the interesting part from the original medical image.Therefore,image segmentation techniques has great practical value for the development of clinical madicine.First,according to the orginal image,using contourlet transform generated multi-resolution image. Then, the original segmentation image has get by apply the Watershed Algorithm. Watershed segmentation method is a mathematical morphology method based on topology theory.Finally,the original segmentation image is mapped to original resolution by apply the inverse contourlet transform.We can get26segment region from the brain CT image through the contourlet transform and the watershed algorithm,while,there are280segment region only if using the watershed algorithm.The results of there two experiments differed fully254segment region.the result is obviously.In this article,the author conduct an expriment about brain CT image.The result show that the method effectively inhibits over-segmentation phenomenon and receive good segmentation result.It has important theoretical and pratical value. |
