Lung segmentation is the first step to identify any lung-related disease. It is an image processing-based process to obtain the boundary of the lung area from thoracic on CT images. To challenge this scenario, advanced diagnosis methods are needed that requires CT scan images of patient. Radiologists need huge amount of time to detect if any person is having lung cancer or not. To help radiologists, several researchers had proposed many computer-aided diagnosis systems to detect lung-related disease at early stages. In the present work, lung segmentation is done in three dimensions. Image processing techniques are applied named thresholding, morphological operation and active contour to achieve this. At first, preprocessing is done to normalize the value and then Otsu thresholding to divide image into two regions. Morphological operation like erosion is applied to eliminate unwanted region. Active contour is applied at last to segment lung in 3D. Here, 15 subjects are taken from LIDC-IDRI. This method has achieved 0.967 Jaccard index and 0.983 Dice similarity coefficient when compared to ground truth. A 3D view of lung segmentation is also shown.
CITATION STYLE
Sahu, S. P., Kamble, B., & Doriya, R. (2020). 3D Lung Segmentation Using Thresholding and Active Contour Method. In Advances in Intelligent Systems and Computing (Vol. 1097, pp. 369–380). Springer. https://doi.org/10.1007/978-981-15-1518-7_31
Mendeley helps you to discover research relevant for your work.