This paper investigates the possibility to discriminate day-night periods based exclusively on morphological characteristics extracted from P and T waves. Daytime values were determined between 13:00 and 15:00 and nighttime values between 1:00 and 3:00. Initially, P and T waves were selected manually and afterwards in an automatic way, employing a new algorithm. After delimiting the position of P and T waves, thirty four features, most of which employed for the first time for this purpose, were extracted using geometrical measures. All these features were examined for both waves. The results revealed a very clear discrimination for the majority of the features, leading to the conclusion that P and T wave morphology is very different during day and night, whilst the correlation coefficient indicated an association and not an arbitrary alteration. Despite the fact that the selected features were rich descriptors, machine learning techniques were also employed to confirm the discrimination capability.
CITATION STYLE
Zavantis, D., Mastora, E., Kontogiannis, P., & Manis, G. (2018). Discrimination between day and night ECG recordings based on the morphology of P and T waves. In IFMBE Proceedings (Vol. 68, pp. 275–279). Springer Verlag. https://doi.org/10.1007/978-981-10-9038-7_51
Mendeley helps you to discover research relevant for your work.