Segmentation of lung from CT using various active contour models. (January 2019)

• Record Type: Journal Article
• Authors: Nithila, Ezhil E.; Kumar, S.S.
• Is Part Of: Biomedical signal processing and control. Volume 47(2019)
• Page Start: 57
• Page End: 62
• Publication Date: 2019-01
• Subjects: Segmentation -- Active contour -- Level set -- Gaussian filtering -- Signed pressure force; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2018.08.008
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: We present a fast, new algorithmto segment the lung from CT images. We place automatically the initial contour to locate the boundary and identify the concave edges. Our method is fully automatic, needing no prior processing or manual intervention. The new algorithm showed excellent result compared with other active contour models. Our method is robust to image quality and easy to diagnosis the disease at earlier stage. Abstract: The aim of the paper is to develop a region based active contour model using variational level set function for segmentation of lung. Ultimately, segmentation of parenchyma is essential for accurate diagnosis of various lung diseases. Among many imaging modalities, Computed Tomography (CT) is a pioneer to most image analysis applications. This work proposes a powerful technique named Selective Binary and Gaussian filtering-new Signed Pressure Force (SBGF-new SPF) function for segmentation of CT lung images. This process detects the external boundary of the lung and effectively stops the contour even at blurry boundaries. The proposed algorithm was compared with four different active contour models. Comparative experiments demonstrate the advantage of proposed method in terms of computation time and accurate segmented lung.