Separating a mixture of quasi-periodic componets

Thumbnail Image
Θεόδωρος, Μπερμπερίδης
Journal Title
Journal ISSN
Volume Title
In many biomedical engineering applications that involve simultaneous recordings of multiple biological signals we want to extract the original signals from the observed mixtures. Due to the quasi-periodic nature of many of these biological processes, such as the heart and lung activity, we inevitable arrive to the problem of Periodic Component Analysis. We will show that this is a very special case of a broad spectrum of problems, what is known for decades as Blind Source Separation (BSS). We developed methods that attempt to solve this problem by taking into account the periodic nature of the signals as well as the realistic limitations when attempting to solve this problem in real life scenarios, specifically in the case of simultaneous recordings of heart and lung acoustic signals , where the nature of signal transmission in the human body requires a more sophisticated model.
Periodic component analysis, Blind source separation