摘要
针对临床辅助诊断的需求,本文提出一种肺部感兴趣区域快速分割算法。该算法首先利用最大类间方差(OTSU)法对图像进行预分割,然后利用区域生长及小面积消除方法剔除干扰信息,并生成掩模图像,最后运用数学形态学方法对模板进行细化,并将原始图像与掩模图像进行数学运算,从而获得肺部感兴趣区域。实验结果表明,本文提出的算法自动化程度高,分割较为准确。
According to the request of clinical diagnostic-assisted,a fast segmentation algorithm approach on lung ROI is proposed in this paper.Firstly,maximum between-cluster variance method is applied to pre-segment the image.Secondly,the method of region growing and small area elimination is applied to get rid of disturb information and generate mask image in the meantime.Lastly,the mathematical morphology method is applied to refine the template image.And arithmetical operation is used to original image and mask image to get lung region of interest.The experimental results indicate that the algorithm in this paper has a high automation degree and accuracy segmentation.
出处
《中国医疗设备》
2011年第3期28-30,共3页
China Medical Devices
关键词
CT图像
最大类间方差
数学形态学
肺部感兴趣区域
CT肺部成像
CT images
maximum between-cluster variance
mathematical morphology
lung region of interest
lung CT imaging
