Semiautomatic detection of cardiac diseases employing dual tree complex wavelet transform

Abstract

Electrocardiogram (ECG) contains lot of information which can be utilized for a mechanism to detect cardiac abnormalities. The ECG signal is too sensitive to various types of noises as it is of low frequency and has weak amplitude, these noises reduce the diagnostic accuracy and may lead to the incorrect decision of the clinician. So, denoising of ECG signal is an essential requirement for an accurate detection of Heart disease. In this paper, a Dual-Tree Complex Wavelet Transform technique (DTCWT) is presented to denoise the noisy ECG signal and to extract the Principal features followed by implementation of Peak Detection Algorithm. The performance is evaluated on the basis of performance metrics and an increase in SNR is achieved using the technique. With the proposed technique, calculated heart rate is in consensus with the gold standard of the various bench mark databases used and accurate heart disease was determined.