Algoritmy pro zpracování signálů v oblasti elektronického monitorování plodu

Abstract

This dissertation focuses on the use of advanced signal processing methods and methods based on artificial intelligence and machine learning in the field of electronic fetal monitoring. Based on the literature review, a hybrid system was designed for the non-invasive fetal electrocardiogram extraction, the electrohysterogram extraction, and the fetal state classification. The system was verified on real data obtained from clinical practice and its effectiveness was evaluated using objective parameters. The proposed approach proved to be effective in terms of accurate fetal R peaks detection, fetal heart rate determination, ST analysis performance, uterine contractions detection, and fetal state classification into three classes (normal, suspect, and pathologic). Due to its reliability, non-invasiveness, and absence of ultrasound waves, non-invasive fetal electrocardiography could replace conventional cardiotocography currently used in clinical practice and, in addition, open the way for the development of a fetal monitoring system meeting the pillars of the so-called Fetal Monitoring 4.0.