Extraction and Statistical Analysis of Fetal Electrocardiogram

This paper proposes a numerical algorithm for extracting fECG from aECG recordings and introduces a statistical description of these biosignals which are generally difficult to extract. This gives a new perspective for analysis of medical signals. Initially, the preprocessing stage of fECG signals is described, and then the fECG extraction using an adaptive algorithm is performed, followed by the application of various statistical methods, used in order to extract significant features and to find out different properties of this biosignal, of clinical interest. Notions like randomness or stationarity are recalled to describe the behavior of fetal electrocardiogram, applying different tests, which improve the extraction of the significant characteristics or the modelling the most common cardiac pathologies, bradycardia and tachycardia.