Lung CT Image Segmentation Based on Border Approximation
Huang Zhiding1 Sun Hong1,2,3*
1 School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology, Shanghai 200093, China 2 Business School, University of Shanghai for Science and Technology, Shanghai 200093, China 3Shanghai Key Lab of Modern Optical System, Shanghai 200093, China
Abstract：Segmentation of the lungs in chest-computed tomography (CT) image performs as an important preprocessing step in Computer-aided detection (CAD). The result brings a great effect for the further analysis and diagnosis. As the intensity of pleural nodule is close to the peripheral lung parenchyma, these lesions are not able to be segmented correctly using traditional method. The aim of this paper is to segment the lung including juxtapleural nodules in order to provide this focus for CAD system for the further analysis. This paper proposed a method that combined the Graham algorithm with border matching approximation to correct the contour of lung, and obtained the mask and multiply it by original image to segment lung image with juxtapleural nodules. Processing the 68 sample CT images including nodules from LIDC (Lung Image Database Consortium) through adopting new method proposed in this paper, and comparing this method with a traditional one, and analyzing the accuracy of this method and the rate of oversegmentation and undersegmentation, the accuracy of 98.5% , the oversegmentation rate of 1.35% and the undersegmentation of 0.51% were determined, which proved the effectiveness of this method.
[1] Mansoor A, Bagci U, Foster B, et al. Segmentation and image analysis of abnormal lungs at CT: Current approaches, challenges, and future trends[J]. Radiographics, 2015, 35(4):1056-1076. [2] Samuel G Armato III, Maryellen L. Giger, computerized detection of pulmonary nodules on CT Scans[J]. Radio Graphics, 1999,19:1303-1311. [3] Guo PX, Guo H, Dai J. A Novel lung segmentation and boundary correction method[C]//Biomedical Engineering and Informatics. Tianjin: IEEE, 2009:1 - 4. [4] Shen S, Bui AAT, Cong J, et al. An automated lung segmentation approach using bidirectional chain codes to improve nodule detection accuracy[J]. Computers in Biology and Medicine, 2015,57:139-149. [5] 李金, 郑冰,梁洪,等.基于改进凸包算法的肺实质分割研究[J]. 中国生物医学工程学报, 2013, 32(4):484-490. [6] Pu J, Roos JC, Napel S, et al. Adaptive border marching algorithm: automatic lung segmentation on chest CT images[J]. Computerized Medical Imaging & Graphics, 2008, 32(6):452-462. [7] Chunming L, Chenyang X, Changfeng G, et al. Distance regularized level set evolution and its application to image segmentation.[J]. IEEE Transactions on Image Processing, 2010, 19(12):3243-3254. [8] Dai S, Lu K, Dong J, et al. A novel approach of lung segmentation on chest CT images using graph cuts[J]. Neurocomputing, 2015, 168:799-807. [9] Ju W, Xiang D, Zhang B, et al. Random walk and graph cut for co-segmentation of lung tumor on PET-CT images[J]. IEEE Transactions on Image Processing, 2015, 24(12):5854-5867. [10] Tasci E, Ugur A. Shape and texture based novel features for automated juxtapleural nodule detection in lung CTs [J].Journal of Medical Systems,2015, 39(5): 1-13. [11] Mohsen K, Zohreh A, Farshad T. Lung nodulesegmentation and recognition using SVM classifier and active contour modeling:A complete intelligent system[J].Computers in Biology and Medicine ,2013,43:287-300. [12] Sensakovic WF, Starkey A, Armato SG. A general method for the identification and repair of concavities in segmented medical images[J]. IEEE Nuclear Science Symposium Conference Record Nuclear Science Symposium. Dresden:IEEE,2008, 20(7):5320-5326.
