Extrakce plodového elektrokardiogramu metodami soft computingu

Abstract

This master’s thesis deals with a proposal of software solution for adaptive fetal electrocardiogram (fECG) extraction by soft computing methods from non-invasive fetal heart monitoring. First part of this thesis is dedicated to a complex overview of soft computing methods used in fECG elicitation. Next part of thesis devotes to proposal of software solution for adaptive fECG extraction. The proposed adaptive system is based on adaptive neuro-fuzzy interference system (ANFIS) with hybrid learning algorithm. The functionality of the proposed system is tested by Signal to noise ratio (SNR), Root Mean Square Error (RMSE) and Percent Root Difference (PRD). Quality of fECG filtration is also evaluated for different setting parameters of the system.