摘要
传统单一模式下的分割方法,无法完成完整序列的CT图像分割。由于肺癌CT图像需结合临床医学知识进行医用调窗处理,不同窗宽、窗位对危及器官肺分割结果会产生影响。基于肺癌放疗计划的危及器官肺分割,提出一种完整序列下危及器官肺实质分割方法。首先引入医用调窗技术对肺癌CT图像进行调窗处理,并对其进行伪影噪声的消减,然后对不同窗的图像采用改进的三维Otsu分割法和区域生长法进行分割,最后将计算机分割出的危及器官肺与临床医师手动分割的结果进行对比,验证分割算法的准确性,为临床危及器官肺分割提供参考。
In traditional single-mode segmentation method cannot segment the whole series of CT images, because the CT image of lung cancer needs to be processed with therapeutic window adjustment by combining with clinical medical knowledge. Both the width and level of window have influences on the segmentation results of the lung at risk. Based on the segmentation in the radiotherapy plan for lung cancer, a new segmentation method for the CT series of pulmonary parenchyma is proposed. Firstly, medical therapeutic window adjustment is introduced to process the CT image of lung cancer and reduce the noise artifacts. Secondly, an improved three-dimensional Otsu segmentation method and region growing algorithm are used to segment the images in different windows. Finally, the results of segmentation by proposed method are compared with clinician manual segmentation. The accuracy of the segment algorithm is verified. The proposed method can be used to make reference for the segmentation of the lung of organ at risk.
出处
《中国医学物理学杂志》
CSCD
2016年第11期1097-1107,共11页
Chinese Journal of Medical Physics
基金
国家科技重大专项(2014ZX04012015)
关键词
肺癌
放疗计划
危及器官肺
医用调窗处理
三维Otsu分割法
区域生长法
lung cancer
radiotherapy plan
lung of organ at risk
therapeutic window adjustment
three-dimensional Otsu segmentation method
region growing algorithm
