Lung Lobe Segmentation Based on HRCT Data

Abstract

Nowadays, the lung lobe segmentation is the most basic step in Lung CAD (Computer-aided diagnosis) and is playing an increasingly important role in the early diagnosis of lung diseases and the analysis of pulmonary functions. The key to achieving lung lobe segmentation is to detect and locate lung fissures. With the wide applications of HRCT (High-Resolution Computed Tomography), CT data with higher contrast can be got, thus making it possible to locate lung fissures more accurately. In this paper, a lung fissure extraction algorithm based on the two-dimensional chest HRCT data is proposed. First, A linear structure enhancement filter based on the Hessian matrix is designed to enhance the contrast of lung fissures; then, according to the idea of Canny operator, ridge of the image is extracted, which allows the location of the fissures to be determined accurately; finally, the Uniform Cost Method is applied to the detection of ridge of the fissures and the extraction of them are achieved. Experiments show that this algorithm can realize the extraction of lung fissures and achieve the lung lobe segmentation with good effects.