Komparativní studie využití adaptivních metod jako součásti hybridního systému pro extrakci plodového EKG signálu

Abstract

The main goal of this diploma thesis is the extraction of the fetal ECG signal exclusively from abdominal signals (aECG) scanned non-invasively by transabdominal method, using electrodes placed on the mother's abdomen. The aECG signal contains not only the fetal but also the maternal component of the ECG (mECG) and other interfering signals. The theoretical part of the work deals with the current state of this issue, a description of methods used to monitor the fetus in clinical practice and the classification of selected methods and their mathematical analysis. Furthermore, the work focuses on the design and implementation of a hybrid system that uses a combination of non-adaptive and adaptive methods to extract the fECG signal. Evaluation of filtration quality is performed on the basis of statistical parameters ACC and F1 and also on the basis of determination of fetal heart rate variability (fHR). Verification of the functionality of the proposed algorithm was performed on real data from clinical practice and also on data measured directly in the laboratory.

Description

Subject(s)

Fetal electrocardiogram, non-invasive fetal monitoring, hybrid algorithm, ECG extraction, Independent Component Analysis (ICA), LMS, NLMS, BLMS, ADALINE (Adaptive Linear Neuron).

Citation