Pulmonary Nodule Segmentation with Modified Variable N-Quoit Filter Combining Border Smoothing and Correction

• Published in: 2015 7th International Conference on Intelligent Human Machine Systems and Cybernetics (IHMSC) Vol. 1; pp. 374 - 377
• Main Authors: Jinke Wang, Yuanzhi Cheng, Quanxu Ge
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2015
• Subjects: adaptive border marching; Cancer; Computed tomography; concave-convex correction; Hospitals; Image segmentation; Lungs; pulmonary nodule segmentation; Smoothing methods; Transforms; variable N-quoit filer
• ISBN: 9781479986453; 1479986453
• DOI: 10.1109/IHMSC.2015.29

In this paper, an automatic pulmonary nodule segmentation scheme is proposed using modified variable N-quoit filter (VNQ), combined with lung boundary smoothing and correction. The whole scheme is mainly divided into three stages: lung parenchyma segmentation, lung boundary smoothing and correction, and candidate nodules segmentation. In the lung parenchyma segmentation stage, an adaptive border marching algorithm (ABM) is implemented for rough outline of the lung parenchyma, In the lung border smoothing and correction stage, arc length based algorithm and concave-convex based correction methods are used to detect the pleural nodules, In the final stage, candidate nodules are obtained by the use of modified variable N-quoit filter and region growing algorithm. We validate our scheme on 10 CT exams by comparing our scheme with traditional variable N-quoit filter, and the preliminary results indicate a good efficiency of the method we proposed.