基于自动选取多个阈值的乳腺X光图像分割方法
New Segmentation Method of Mammary X-ray Images Based on Finding the Optimum Thresholds
采用分水岭算法对数字X光图像的直方图一维信号进行处理。首先计算出图像的直方图，然后对直方图进行形态学滤波，提取出直方图的包络，在此基础上对直方图的包络采用分水岭算法自动提取出多个最佳阈值。用提取出的阈值进行基于阈值的初步分割，然后进一步采用基于标记的分水岭算法对初步分割的二值图像进行处理。通过实验表明，采用本文提出的算法可以有效分割出乳腺X光图像中的感兴趣区域。
In this paper, we present a new method to segment X-ray images. First, the histogram of image is calculated, and then the histogram is filtered by mathematical morphology to extract the envelop curve. At last, the thresholds are calculated by watershed algorithm according to the envelop curve. Then on the basis of the first segmentation using calculated threshold, X-ray image is segmented based on marked watershed algorithm. Experiments prove this method is effective to segment interest region.
李哲、黄廉卿
医学研究方法基础医学
直方图最佳阈值分水岭算法X光图像
HistogramOptimum thresholdWatershed algorithmX-ray image
李哲,黄廉卿.基于自动选取多个阈值的乳腺X光图像分割方法[EB/OL].(2005-06-27)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/200506-144.点此复制
评论