Estimação paramétrica para um modelo de sinais de eletrocardiograma sintético utilizando o algoritmo da região de confiança

Abstract

The application of computer models in methods for detecting heart diseases has become very important since these diseases are the leading cause of mortality worldwide. There are some mechanisms for diagnosing such heart diseases, among them the Electrocardiogram (ECG), whose fundamental study brought to light studies that proposed to develop synthetic signals of ECGs. The present study presents a methodology capable of giving ECG signals through parametric estimation of a synthetic model. To this end, it was necessary to learn specific concepts about the heart and the electrocardiogram morphology, characterized by the presence of P, Q, R, S, and T waves. It was noted that the combination of different values for parameters θi , ai, and bi , which formulate such waves, result in different morphologies for the generated ECG signal. Random values were used to analyze the convergence between the estimated ECG signal and the signal generated from previous works’ data. The main results were obtained by analyzing each parameter and the resulting signal, which showed excellent efficiency.