Method and system for processing a phonocardiographic signal characterizing fetal breathing movement

A system for processing a phonocardiographic signal characterising fetal breathing movement, comprising a phonocardiographic sensor for recording the phonocardiographic signal on an abdominal wall of a mother pregnant with a fetus, and, being in interconnection therewith, a computer that comprises:

• an initial start point determination module comprising
• a filter device for generating 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
• a start point search module for determining a first identified start point corresponding to a fetal breathing movement episode in a frequency subband corresponding to the first identified start point, the start point search module being adapted for searching, investigating presumed fetal breathing movement episodes in the second bandpass-filtered signals for each frequency subband, 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
• an episode discover module for searching, by means of the start point search module, a further start point in an episode search time period starting from the first identified start point, and the episode discover module
• if finds a start point 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 merges it with the first identified start point,
• if finds a start point 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 identifies such a start point being closest to the first identified start point as a second identified start point, and episode validity determination module is applied for the first identified start point and the second identified start point, and after that the episode discover module is applied from the second identified start point in case processing of the phonocardiographic signal is continued,
• if does not find a start point within the episode search time period, then the start point search module is applied for determining another first identified start point in case processing of the phonocardiographic signal is continued.

A method for processing a phonocardiographic signal characterising fetal breathing movement, comprising:

• receiving, by a computer, the phonocardiographic signal from an acoustic sensor;
• applying, by the computer, a first bandpass filter to the phonocardiographic signal to generate a first set of band-pass filtered signals of a main investigation frequency band;
• applying, by the computer, a second bandpass filter to the first set of band-pass filtered signals to divide the main investigation frequency band into a second set of band-pass filtered signals corresponding to multiple frequency subbands;
• analyzing, by the computer and for each frequency subband, the second set of band-pass filtered signals to identify a first start point of a fetal breathing episode, wherein identifying the first start point includes:
• calculating measured movement features for the fetal breathing episode,
• computing an acceptance sum by applying weights to the movement features,
• comparing the acceptance with acceptance threshold values corresponding to the frequency subbands, and
• accepting or rejecting the first start point based on said comparing;
• and
• further analyzing, by the computer and for each subband, a time window beginning after the first start point to identify a further start point of a new fetal breathing episode, wherein identifying the further start point includes:
• calculating measured movement features for the new fetal breathing episode,
• computing another acceptance sum by applying the weights to the measured movement features for the new fetal breathing episode,
• comparing the other acceptance sum with the acceptance threshold values corresponding to the frequency subbands, and
• in response to the further start point being found in a particular subband that is different from a subband corresponding to the first start point at a smaller distance from the first start point than a clustering threshold value, merge the further start point with the first start point,
• in response to the further start point being found in any frequency subband at a larger distance from the start point than the clustering value threshold, or in the frequency subband corresponding to the first start point at a smaller distance from the first start point than the clustering threshold value, then such a start point being closest to the first start point is identified as a second start point, or
• in response to the further start point not being found within the time window beginning after the first start point, identifying yet another start point of a yet another fetal breathing episode.