Classification of Benign/Malignant PNGGOs using K-means algorithm in MDCT Images: A Preliminary Study

Abstract

Lung cancer is one of the most prevalent diseases in the world. Recently, PNGGOs (Pure nodular ground-glass opacity) have been reported to increasing aspect for all CT-detected pulmonary nodules. Moreover, the malignancy rate of PNGGOs is a considerable proportion of benign diseases. In this study, we have developed a computerized classification scheme of PNGGOs malignancy. Segmentation of PNGGOs was performed semi-automatically. After that, the histogram based statistical features and region based features of benign and malignant GGO was extracted. Finally, K-means classifier was applied. Experiment was performed employing 12 CT image sets and 91.67% of accuracy was achieved.