| CPC A61B 8/02 (2013.01) [A61B 5/344 (2021.01)] | 15 Claims |
|
1. A method for processing a phonocardiographic signal characterising fetal breathing movement, in course of which:
in course of an initial start point determination step
in a frequency filtering step, generating, by a computer, respective second bandpass-filtered signals corresponding to frequency subbands having respective subband bandwidths by means of a second bandpass filter from a first bandpass-filtered signal obtained by filtering the phonocardiographic signal to a main investigation frequency band by means of a first bandpass filter, and
determining, by the computer, a first identified start point corresponding to a fetal breathing movement episode in a frequency subband corresponding to the first identified start point in a start point search step, in course of which, investigating presumed fetal breathing movement episodes in the second bandpass-filtered signals for each frequency subband, searching a start point of a fetal breathing movement episode by comparing an acceptance sum obtained by weighting measured movement features obtained for the presumed fetal breathing movement episodes with respective acceptance threshold values corresponding to each of the frequency subbands, and
in an episode discovering step, searching, by the computer, a further start point applying the start point search step in an episode search time period starting from the first identified start point, and
if a start point is found in a frequency subband other than the subband corresponding to the first identified start point at a smaller distance from the first identified start point than a clustering threshold value, then it is merged with the first identified start point,
if a start point is found in any frequency subband at a larger distance from the identified start point than the clustering threshold value, or in the frequency subband corresponding to the first identified start point at a smaller distance from the first identified start point than the clustering threshold value, then such a start point being closest to the first identified start point is identified as a second identified start point, and an episode validity determination step is performed for the first identified start point and the second identified start point, and after that continuing the method with the episode discovering step from the second identified start point in case processing of the phonocardiographic signal is continued,
if a start point is not found within the episode search time period, then continuing the method with a step adapted for determining another first identified start point in case processing of the phonocardiographic signal is continued.
|