ECG signal features extraction

The electrocardiogram (ECG) signal is used to assess electrical abnormalities and provides vital information about of heart health. One problem in ECG analysis is the feature extraction due to the intrinsic noise. This paper presents a ECG feature extraction method that consists of the morphology analysis, the fiducial point localization, and time intervals measurements by using both continuous and discrete wavelets transform (CWT and DWT, respectively). From ECG signals obtained from CardioExpress SL3 medical device, we obtained the principal features by using the Wavelet theory. The results were validated and compared with the established values by the International Committees, the Common Standards for quantitative Electrocardiography (CSE), with a Sensitivity (Se) of 96% and with a standard deviation (std) less than 8%.