A Zernike Moment based methodology for heart disease detection

Abstract

This paper details a Zernike Moments and Fuzzy-C-Means clustering based technique to identify the nature of an ECG image. The proposed method can detect whether the ECG image belongs to a normal heart or a diseased heart. In the second case it can indicate the disease of the heart also. The method has been tested on four databases- congestive heart failure database, ventricular tachyarrhythmia database, atrial fibrillation database and normal sinus rhythm database. The experiment shows that the proposed technique is successful in 98.7% cases.