Comparative study of feature classification methods for mass lesion recognition in digitized mammograms
Giovanni Luca Masala, Sabina Tangaro, Bruno Golosio, P. Oliva, Simone Stumbo, R. Bellotti, Francesco De Carlo, Gianfranco Gargano, Donato Cascio, Francesco Fauci, Rosario Magro, G. Raso, U. Bottigli, A. Chincarini, I. De Mitri, Giorgio De Nunzio, Ilaria Gori, Alessandra Retico, P. Cerello, Sorin Cristian Cheran · Kent Academic Repository (University of Kent) · 2007
In this work a comparison of different classification methods for the identification of mass lesions in digitized mammograms is performed. These methods, used in order to develop Computer Aided Detection (CAD) systems, have been implemented in the framework of the MAGIC-5 Collaboration. The system for identification of mass lesions is based on a three-step procedure: a) preprocessing and segmentation, b) region of interest (ROI) searching, c) feature extraction and classification. It was tested on a very large mammographic database (3369 mammographic images from 967 patients). Each ROI is characterized by eight features extracted from a co-occurrence matrix containing spatial statistics information on the ROI pixel grey tones. The reduction of false-positive cases is performed using a classification system. The classification systems we compared are: Multi Layer Perceptron (MLP), Probabilistic Neural Network (PNN), Radial Basis Function Network (RBF) and K-Nearest Neighbours classifiers (KNN). The results in terms of sensitivity (percentage of pathological ROIs correctly classified) and specificity (percentage of nonpathological ROIs correctly classified) are presented. MLP and RBF outperform other classification algorithms by about 8% of the area under the ROC curve