Classification of lung nodules and arteries in computed tomography scan image using principle component analysis

Abstract

There is still a lack of a good method of diagnosing pulmonary nodules in CT Scan automatically, causing medical staff to observe a 2-D CT Scan data manually and interpreting data one by one. This procedure is course less effective. In addition, lung specialists may differ in determining pulmonary nodules. The purpose of this research is to classify pulmonary nodules and artery automatically on chest Ct Scan image using Principle Component Analysis (PCA). This study includes 3 steps. The first is lung organ segmentation using Active Appearance Model (AAM). The second step is segmentation of candidate nodules using morphological math. While the last step is classification of pulmonary nodules and artery using Principle Component Analysis method. The output from classification process is image of nodule and artery. Results of testing, obtained the performance of classification system accuracy is 90%.