Interpretation of normal and pathological beats using multiresolution wavelet analysis

The discrete wavelet transform has great capability to analyze the temporal and spectral properties of non stationary signal like electrocardiogram (ECG). In this paper, we developed and evaluated a robust algorithm using multiresolution analysis based on the discrete wavelet transform (DWT) for twelve-lead ECG temporal feature extraction. The study, with support of physiological knowledge, attempted interpretation of ECG beats with different patterns. Selection of appropriate group of wavelet coefficients along with decision rules is used to determine P, Q, R, S and T wave locations, amplitudes, onsets and offsets We evaluated the algorithm on normal and abnormal beats from various manually annotated databases from physiobank with different sampling frequency. An appropriate value of threshold at QRS detector offered sensitivity of 99.5% and positive predictivity of 98.9% over the first lead of the MIT-BIH Arrhythmia Database.