Automated heart abnormality detection using sparse linear classifiers.

In this article, the task of building a computer-aided diagnosis system that can automatically detect wall-motion abnormalities from echocardiograms was addressed. Some medical background on cardiac ultrasound and the standard methodology used by cardiologists to score wall-motion abnormalities were provided. Real-life dataset, which consists of echocardiograms used by cardiologists at St. Francis Heart Hospital to diagnose wall-motion abnormalities were also described. The paper provides an overview of the proposed system, which was built on top of an algorithm that detects and tracks the inner and outer walls of the heart. It consists of a classifier that classifies the local region of the heart wall (and the entire heart) as normal or abnormal based on the wall motion. A methodology for feature selection and classification, followed by our experimental results was also described. The novel feature selection technique results in a robust hyperplane-based classifier that achieves the best performance in terms of AUC (area under the curve) and number of features selected when compared to three other well-known classification algorithms