Automatic Detection, Classification and Measurement of Lung Nodules using Combined Classifiers in CT Scan Images

Lung cancer is one of the hardest and most dangerous types of known cancer in the world which can be detected in its beginning stages as a small mass of tissue, less than 3 cm in diameter, called a nodule. These nodules are classified to two classes of benign or malignant. In this paper, a detection system for detection and classification of lung nodules is proposed which in the first phase, lungs are separated from the CT scan images according to the active contour segmentation method. Next, based on the SIFT features the proposed Bagging classifier, classifies the lung images into two classes of patient and healthy. In the second phase, according to a fully automatic Graph-Cut segmentation method the nodules are extracted from patient images and their diameters are measured. Finally, nodules are classified to two classes of benign and malignant based on their size and texture Haralick features. To evaluate the proposed method, images of the LIDC database are used and its performance in detection of nodules compared to other methods has an accuracy of 97% and in classification of nodules to benign and malignant an accuracy of 96% is reached.