血管粘连型肺结节图像的序列分割方法  

Sequence segmentation method for juxta-vascular lung nodule images

作　　者：张伟[1] 张小龙[2] 赵涓涓[1] 强彦[1] 唐笑先[3] ZHANG Wei1 , ZHANG Xiao- long2 , ZHAO Juan- juan1 , QIANG Yan1, TANG Xiao-xian3(1. College of Computer Science and Technology, Taiyuan University of Technology, Jinzhong 030600, China; 2. College of information Science and Technology, Pennsylvania State University, University Park 16802, USA; 3. Department of PET/CT Center, Shanxi Provincial People＇s Hospital, Taiyuan 030024, Chin)

机构地区：[1]太原理工大学计算机科学与技术学院,山西晋中030600 [2]宾夕法尼亚州立大学信息科学与技术学院,宾西法尼亚州尤尼弗西蒂帕克16802 [3]山西省人民医院PET/CT中心,山西太原030024

出　　处：《计算机工程与设计》2018年第8期2550-2556,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61373100);虚拟现实技术与系统国家重点实验室开放基金项目(BUAA-VR-17KF-14;BUAA-VR-17KF-15);山西省回国留学人员科研基金项目(2016-038)

摘　　要：为解决以往分割算法对血管粘连型结节分割不准确以及分割效率较低等问题,提出基于超像素和稀疏子空间聚类的序列肺结节图像分割方法。对CT图像进行序列肺实质分割,提取感兴趣图像序列,采用改进的超像素序列分割方法对感兴趣图像序列进行过分割,对所有的超像素样本提取新特征,包括对比度增强直方图特征、超像素样本邻域纹理特征以及基于先验知识的位置信息特征,采用距离约束稀疏子空间聚类算法对超像素样本进行聚类,得到序列肺结节掩膜,最终得到序列肺结节图像。实验结果表明,该方法能准确高效地分割序列血管粘连型结节图像。To address the problems that existing segmentation algorithms cannot accurately segment juxta-vascular nodules and have poor efficiency,a sequence segmentation method based on superpixels and sparse sub-space clustering was presented.The lung parenchyma image sequences of CT images were segmented.The regions of interest（ROIs）were obtained.The ROIs image sequences,were then segmented using the improved superpixel sequence segmentation method.The new features of all the superpixel samples were extracted,including contrast enhancement histogram features,the texture features of the neighborhood of the superpixel samples and the location information features based on prior knowledge.The distance-constrained sparse subspace clustering algorithm was utilized to perform superpixel samples clustering to obtain lung nodule mask sequences.The lung nodule image sequences were then obtained.Experimental results show that the presented method can accurately and efficiently segment the juxta-vascular nodule image sequences.

关 键 词：序列分割 血管粘连型结节 超像素 特征提取 稀疏子空间聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]