Although a significant bibliography exists on the application of signal processing methods to ECG signals, the optimal configuration of these methods so as to maximize their performance on clinical data is a complex problem that is seldom covered in the literature. This is particularly the case for the signal processing chains proposed for the detection and segmentation of individual beats, which are often characterized by a significant number or parameters (filter cut-off frequencies, thresholds, etc.). In this chapter we propose an automated method, based on evolutionary computing, to optimize these parameters in a joint fashion. A brief state of the art on current ECG segmentation methods is presented and a complete signal processing chain, adapted to the detection and segmentation of ECG signals is proposed. The evolutionary optimization method is described and applied to two different monitoring applications: the detection of myocardial ischemia episodes on adult patients and the characterization of apnea-bradycardia events on preterm infants.
CITATION STYLE
Hernández, A. I., Dumont, J., Altuve, M., Beuchée, A., & Carrault, G. (2014). Evolutionary optimization of ECG feature extraction methods: Applications to the monitoring of adult myocardial ischemia and neonatal apnea bradycardia events. In ECG Signal Processing, Classification and Interpretation: A Comprehensive Framework of Computational Intelligence (Vol. 9780857298683, pp. 237–273). Springer-Verlag London Ltd. https://doi.org/10.1007/978-0-85729-868-3_11
Mendeley helps you to discover research relevant for your work.