Feature Extraction from an ECG Signal of Various Cardiac Patients Using Daubechies Decomposition Technique

Electrocardiograms (ECG) are cardiac signals formed by the cyclical electrical activity of the heart muscles. The signal is very important for cardiovascular disease assessment. The objective of the work can be laid down as ECG beat detection and Various Feature Extractions which includes P- height, Q- height, R- height, S- height, T- height, QRS width. The QRS complex is one of the most important characteristics and identification of R peaks makes the detection of other characteristic peaks efficient and easy. Using first derivative of ECG signal by computing sample to sample differentiation, position of R-peak can easily be detected. After detection of R-peak, corresponding Q-peak, S-peak, P-wave, T-wave were identified by Daubechies Wavelet Decomposition technique. From the designed algorithm we retrieved data sets for different data entered. We have tested the algorithm with different signals of different cardiac diseases and obtained datasets of the various features and plotted it in a graph to obtain the probable ranges for various features of various diseases. For this work we have used short duration ECG data from the physionet MIT-db database. The beats were segmented based on foot detection and the fiducial points were extracted. Based on the fiducial points, beat wise features were extracted which were fed to the designed algorithm.