An Accurate and Generalized Approach to Plaque Characterization in 346 Carotid Ultrasound Scans

Computer-aided diagnosis (CAD) of carotid atherosclerosis into symptomatic or asymptomatic is useful in the analysis of cardiac health. This paper describes a patented CAD system called Atheromatic (TM) for symptomatic versus asymptomatic plaque classification in carotid ultrasound images. The system involves two steps: 1) feature extraction using a combination of discrete wavelet transform and averaging algorithms and 2) classification using a support vector machine (SVM) classifier for automated decision making. The CAD system was evaluated using a database consisting of 150 asymptomatic and 196 symptomatic plaque regions which were labeled using the ground truth based on the presence or absence of symptoms. Threefold cross-validation protocol was adapted for developing and testing the classifiers. We observed that the SVM classifier with a polynomial kernel of order 2 was to achieve a classification accuracy of 83.7%.