how to detect unborn child’s heartbeat without interventions? or why digital signal processing is cool
electrical signals are measured from a pregnant’s mother’s adomen and thorax and the task is to extract a trace of the unborn child’s electrocardial (heartbeat) signal (ECG signal).
this is a nontrivial and clinically relevant task, and all sorts of basic and advanced signal processing and machine learning methods have been investigated for tackling it.
i’m trying to solve this task while reading the Digital Filters book; applying methods that I see there so that they stick. i wanted to share some nice results, but so far, i couldn’t get it to work((
the idea is, one is given a few signals which are all dominated by one source, and a few of them have a faint signal that one wants to bring to the foreground, by removing the dominant. this task is called noise cancellation (here the “noise” is the mother’s ECG).
on the last image, there are my professor’s results, and imho, they look too good to be true