摘要
基于肺部CT图像灰度分布的特征,提出一种快速而有效的肺实质分割方法。该方法将图像阈值化和快速区域填充方法结合起来,可以自动将肺实质区域从肺部CT图像中分割出来。它首先通过一个预处理步骤以滤除随机噪声,然后采用阈值化方法对图像进行二值化,接着对所得二值图像进行腐蚀操作和面积滤波处理,并运用一种新奇而简单的区域填充方法对二值图像中非肺部区域进行填充以滤除各种干扰区域,最后经简单处理后即可从原始CT图像中分割出肺部区域。该方法实现简单高效,最终的实验结果证明了它的有效性。
Based on grey distribution characteristic of pixel intensity in lung CT images, a fast and efficient pulmonary parenchyma segmen- tation method is introduced in the paper. Associating image thresholding approach with fast flood filling technique, this method can segment pulmonary parenchyma from pulmonary CT image automatically. After a preprocessing step taken for filtering random noise, it makes binarisation on the image slice utilising thresholding method,and after an erosion operation made together with an area-fihering process on the obtained binary image, the method applies just a novel and simple flood filling to fill the non-lung area in binary image to filter out various interference sections, and at last the lung area can be segmented from the primitive CT slice with an easy treatment. The presented method is easy and efficient to be implemented. The final experiment results have proved its validity.
出处
《计算机应用与软件》
CSCD
2010年第3期78-79,113,共3页
Computer Applications and Software
基金
上海高校选拔培养优秀青年教师科研专项基金(358536)
上海市重点学科建设项目(P0502)
关键词
计算机辅助检测
肺实质分割
区域填充
腐蚀
Computer-aided detection Pulmonary parenchyma segmentation Flood filling Erosion
