首页>
外国专利>
ventricular automatic segmentation method using edge classification and region growing techniques
ventricular automatic segmentation method using edge classification and region growing techniques
展开▼
机译:利用边缘分类和区域生长技术的心室自动分割方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
The present invention can be processed more accurately divide the ventricle automatically from the reduced magnetic resonance imaging of the heart, as in consideration of the influence on the part of the voxels of cattle ventricular edge classification and automatic segmentation method using the region growing method that can calculate an accurate blood flow, extract the seed point from the input image and the heart chamber, the seed point extraction with respect to the input image by performing a region growing from seed point partitioning algorithm that is extracted from the process, an initial extraction of the ventricular region, with respect to the input image by using the initial heart chamber, by removing the gray level distortion compensation and noise, calculating a statistic value of the brightness signal and, after creating a polar coordinate image from the distortion compensation, and the noise removed image by detecting an edge, the edge information by classification, and then estimates the mean value of the brightness value of the myocardium, using the brightness average value for the estimated cardiac and, finally, to divide the ventricular region from the input image through the expansion region segmentation algorithm, also, defining a brightness weighting function for taking into account the calculated blood flow of small voxel, and accurately calculate the ventricular blood flow by using the weight function The.
展开▼