A level set based method for lung segmentation in CT images
Shiva Azimi, Hossein Rabbani · 2014
In this paper an automatic computer-aided (CAD) method is utilized for lung segmentation using computed tomography (CT) images. We segmented lung regions - based on the CT data- with nodules attached to the chest wall by using level set modeling. This method is made up of 3 steps: In the first step, an adaptive fuzzy thresholding operation is used to binarize the CT images; in the second step, the lung with non-isolated nodules is segmented applying both level set modeling and convex hull algorithm. In the third step, by using the shape features of lung lobe, the lung is segmented. The experimental results show an accuracy of 98% by our method with out performance other exiting methods.