This paper gives a comparative analysis of methods for classifying pulmonary nodules using computer-tomography images and a comparative evaluation of the information content of the attributes that are used, as well as an estimate of the effectiveness of classifying pulmonary nodules using various machine-learning algorithms. Problems involving the visual classification of pulmonary nodules and conditions for improving its accuracy are investigated. Sets of the most informative attributes are chosen for classifying pulmonary nodules. The accuracies with which pulmonary nodules are classified into benign and malignant are obtained.
© 2017 Optical Society of AmericaPDF Article