基于连续最大流的三维肺实质快速分割算法

doi:10.12068/j.issn.1005-3026.2020.04.003

东北大学学报:自然科学版 ›› 2020, Vol. 41 ›› Issue (4): 470-474.DOI: 10.12068/j.issn.1005-3026.2020.04.003

基于连续最大流的三维肺实质快速分割算法

赵海，周冰玲，朱宏博，窦圣昶

(东北大学计算机科学与工程学院，辽宁沈阳110169)

收稿日期:2019-05-28 修回日期:2019-05-28 出版日期:2020-04-15 发布日期:2020-04-17
通讯作者: 赵海
作者简介:赵海(1959-)，男，辽宁沈阳人，东北大学教授，博士生导师.
基金资助:
中央高校基本科研业务费专项资金资助项目(N161608001)；国家社会科学基金资助项目(18ZD23).

Fast Segmentation Algorithm of 3D Lung Parenchyma Based on Continuous Max-Flow

ZHAO Hai， ZHOU Bing-ling， ZHU Hong-bo， DOU Sheng-chang

School of Computer Science & Engineering， Northeastern University， Shenyang 110169， China.

Received:2019-05-28 Revised:2019-05-28 Online:2020-04-15 Published:2020-04-17
Contact: ZHOU Bing-ling
About author:-
Supported by:
-

摘要/Abstract

摘要： 肺结节是肺癌的表征形式，形状结构多样且易与正常组织产生粘连，使分割存在困难.提出了一种基于空间约束的三维肺实质分割算法，实现对肺实质组织的分割及目标区域的获取.首先使用SLIC方法将二维CT序列图像构建成超像素图像矩阵，并对矩阵进行稀疏化处理，降低矩阵维度.然后连接相邻切片间的超像素构造肺实质组织的三维结构.最后采用连续最大流方法对构造的三维肺部结构进行分割.实验结果表明，所提算法能够快速准确地分割三维肺实质组织，对不同类型肺结节的分割均取得较好结果，具有一定的临床应用价值.

关键词: 肺实质分割, 空间约束, 连续最大流, 能量函数, 矩阵稀疏化

Abstract: Pulmonary nodules are the main form of lung cancer， which has complex shape and structure and is easy to adhere to normal tissues， making it difficult to segment. A three-dimensional lung parenchyma segmentation method based on spatial constraints was proposed to achieve segmentation of lung parenchyma tissue and acquisition of target regions.First， the SLIC method was used to construct a two-dimensional CT sequence image into a superpixel image matrix， and the matrix was thinned to reduce the dimension of the matrix. Then the nodes between adjacent slices were connected to construct a three-dimensional structure of the lung parenchyma. Finally， the continuous max-flow method was used to segment the constructed three-dimensional lung structure. The experimental results showed that the proposed algorithm can quickly and accurately segment three-dimensional lung parenchyma tissue， and obtain good results for segmentation of different types of lung nodules， which has certain clinical application value.

Key words: segmentation of lung parenchyma, space constraint, continuous max-flow, energy function, sparse matrix

中图分类号:

TP391

赵海，周冰玲，朱宏博，窦圣昶. 基于连续最大流的三维肺实质快速分割算法[J]. 东北大学学报:自然科学版, 2020, 41(4): 470-474.

ZHAO Hai， ZHOU Bing-ling， ZHU Hong-bo， DOU Sheng-chang. Fast Segmentation Algorithm of 3D Lung Parenchyma Based on Continuous Max-Flow[J]. Journal of Northeastern University Natural Science, 2020, 41(4): 470-474.

参考文献

[1]Zhu W，Liu C，Fan W，et al.Deeplung:3D deep convolutional nets for automated pulmonary nodule detection and classification［C］//IEEE Winter Conference on Applications of Computer Vision(WACV).Los Angeles，2018:673-681.
[2]Dou Q，Chen H，Yu L，et al.Multilevel contextual 3D CNNs for false positive reduction in pulmonary nodule detection［J］.IEEE Transactions on Biomedical Engineering，2016，64(7):1558-1567.
[3]Jin H，Li Z，Tong R，et al.A deep 3D residual CNN for false positive reduction in pulmonary nodule detection［J］.Medical Physics，2018，45(5):2097-2107.
[4]依玉峰，高立群，郭丽.一种基于改进随机游走的肺结节分割方法［J］.东北大学学报(自然科学版)，2012，33(3):17-21.(Yi Yu-feng，Gao Li-qun，Guo Li.A lung nodule segmentation method based on improved random walk［J］.Journal of Northeastern University(Natural Science)，2012，33(3):17-21.)
[5]Ren X，Malik J.Learning a classification model for segmentation［C］//International Conference on Computer Vision.Nice，2003:10-17.
[6]Achanta R，Shaji A，Smith K，et al.SLIC superpixels compared to state-of-the-art superpixel methods［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2012，34(11):2274-2282.
[7]Nie W Z，Liu A A，Gao Z，et al.Clique-graph matching by preserving global & local structure［C］//IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Boston，2015:4503-4510.
[8]Ford L R，Fulkerson D R.Maximal flow through a network［J］.Canadian Journal of Mathematics，1956，8:399-404.
[9]Yuan J，Bae E，Tai X C.A study on continuous max-flow and min-cut approaches［C］//Computer Vision & Pattern Recognition.San Francisco:IEEE，2010:13-18.
[10]Simon N，Friedman J，Hastie T，et al.A sparse-group lasso［J］.Journal of Computational & Graphical Statistics，2013，22(2):231-245.
[11]刘建伟，崔立鹏，刘泽宇，等.正则化稀疏模型［J］.计算机学报，2015(7):1307-1325.(Liu Jian-wei，Cui Li-peng，Liu Ze-yu，et al.Regularized sparse model［J］.Chinese Journal of Computers，2015(7):1307-1325.)
[12]Taha A A，Hanbury A.Metrics for evaluating 3D medical image segmentation:analysis，selection，and tool［J］.BMC Medical Imaging，2015，15:29-56.
[15]关守平，房少纯.一种新型的区间-粒子群优化算法［J］.东北大学学报(自然科学版)，2012，33(10):1381-1384.(Guan Shou-ping，Fang Shao-chun.A new interval particle swarm optimization algorithm［J］.Journal of Northeastern University(Natural Science)，2012，33(10):1381-1384.)

基于连续最大流的三维肺实质快速分割算法

Fast Segmentation Algorithm of 3D Lung Parenchyma Based on Continuous Max-Flow

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 4

编辑推荐

Metrics

本文评价

[1]	王旭，张鑫慧，杨丹. MIT中灵敏度矩阵的聚类优化[J]. 东北大学学报:自然科学版, 2019, 40(10): 1369-1375.
[2]	贾同;孟琭;赵大哲;王旭;. 基于CT图像的自动肺实质分割方法[J]. 东北大学学报(自然科学版), 2008, 29(7): 965-967+975.
[3]	片兆宇;高立群;郭丽;王坤;. 多阶段边缘检测算法[J]. 东北大学学报(自然科学版), 2008, 29(5): 637-640.
[4]	郭薇;魏颖;周翰逊;薛定宇;. 基于混合分类的肺结节检测算法[J]. 东北大学学报(自然科学版), 2008, 29(11): 1528-1531.