摘要
在对心脏CT图像左心室区域的分割过程中,常常需要从图像背景中分离出心脏中各个腔体的图像,以便获取腔体轮廓特征。本文中利用图像灰度直方图与高斯概率分布的相似性,应用高斯分布模拟灰度图像直方图方法,求解出最优阈值完整的分离出心脏的各个腔体区域。该方法充分考虑了图像背景区域与目标区域灰度交错以及心脏CT图像中造影剂分布不均匀的情况,分割结果中,目标区域轮廓清晰,准确。
In the process of identification and segmentation of CT images of the left ventricle in the heart region,it is often necessary to isolate the background from the image the heartq in each cavity of the images in order to obtain characteristics of cavity contour.In this paper,by using the similarity between the image histogram and the Gaussian probability distribution,we present a method that comes from the thought of Gaussian distribution of simulating gray image histogram.The optimal threshold value is solved.Each cavity of the heart is isolated completely.The method takes full consideration of the possibility of gray intersect between the image background region and the target area.In Segmentation results,the target area outlines are clear and accurate.
出处
《微计算机信息》
2010年第32期185-187,共3页
Control & Automation
关键词
心脏CT图像
阈值
全局阈值
灰度直方图
高斯概率分布
Cardiac CT Images
Threshold
Global Threshold
Histogram
Gaussian probability distribution
