A new method for ECG signal feature extraction

The paper concentrates on electrocardiographic (ECG) signal analysis for the sake of acquiring proper heartbeat datasets for further detection and recognition of anomalies in an ECG signal. The analysis is based on the distribution of the voltage extreme values of the signal and the time distribution of proper extreme values. The main assumption of the work is creation of as efficient and fast algorithm as possible. The main part of the analysis is carried out on the smoothened signal with the original signal used as the reference for the purpose of keeping desired precision of proper heartbeat identification in the situations when during smoothening necessary data is lost. The proposed method of feature extraction will be used to create the distinctive normal heartbeat samples of patients and the average pattern of the patients' heartbeat for reference purposes in analysis of the real time signal obtained from 1-lead holter carried by a patient constantly.