ECG signal parameters extraction using intelligent adaptive algorithm

摘要

The investigation of standard electrocardiogram signal is one of the basic routine tests, which are used in order to detect or predict various cardiac disorders by interpreting the information provided by the Electrocardiogram. ECG signals are generally affected by various noises such as power line interference and the baseline drift that can be degraded necessary information. The extraction of pure cardio logical indices from noisy measurements has been one of the major concerns of biomedical signal processing. Therefore, an efficient process with good performance (accuracy, speed) must be essential. This signal first undergoes a preprocessing stage for noise cancellation followed by a P-QRS-T waves detection stage. The first stage deals with denoising ECG signal using different types of wavelets. The performances of algorithms applied for biomedical data processing are evaluated and compared in terms of basic parameters such as Mean Square Error (MSE), and Signal to Noise Ratio (SNR). In terms of P-QRS-T waves detection, we have described a comparative study between Pan Tompkins and wavelet algorithms. After applying two P-QRS-T waves detection algorithm on various recordings of MIT-BIH database, it is necessary to validate their efficiency. Several validation tests are bestowed, among them, we are used the sensitivity Se, the positive predictive value P + and the error rate Te (%) in the aim to reach the most efficient technique among the applied algorithms for biomedical data extraction.