Diagnosis of lung nodule using the semivariogram function

Abstract

This paper proposes using the semivariogram function, to help characterize lung nodules as malignant or benign in computerized tomography images. The tests described in this paper were carried out using a sample of 36 nodules, 29 benign and 7 malignant. Fisher's Linear Discriminant Analysis (FLDA), Multilayer Perceptron (MLP) and Support Vector Machine (SVM) were performed to evaluate the ability of these features to predict the classification for each nodule. A leave-one-out procedure was performed to provide a less biased estimate of the classifiers performance. All analyzed classifers have value area under ROC curve above 0.9, which means that the results have excellent accuracy. The preliminary results of this approach are very promising in characterizing nodules using semivariogram function. © Springer-Verlag 2004.