MULTI-CLUSTER SUPPORT VECTOR MACHINE CLASSIFIER FOR THE CLASSIFICATION OF SUSPICIOUS AREAS IN DIGITAL MAMMOGRAMS

摘要

This paper presents a novel technique for the classification of suspicious areas in digitalnmammograms. The proposed technique is based on clustering of input data into numerousnclusters and amalgamating them with a Support Vector Machine (SVM) classifier.nThe technique is called multi-cluster support vector machine (MCSVM) and is designednto provide a fast converging technique with good generalization abilities leading to annimproved classification as a benign or malignant class. The proposed MCSVM techniquenhas been evaluated on data from the Digital Database of Screening Mammographyn(DDSM) benchmark database. The experimental results showed that the proposednMCSVM classifier achieves better results than standard SVM. A paired t-test and Anovananalysis showed that the results are statistically significant