Systems, apparatuses and methods for sensing fetal activity

• US 10,213,120 B2
• Filed: 02/22/2017
• Issued: 02/26/2019
• Est. Priority Date: 03/16/2015
• Status: Active Grant

• Current Assignee: Nuvo Group Ltd.
• Original Assignee: Nuvo Group Ltd.
• Inventors: Oz, Oren, Kam, Amit, Intrator, Nathan
• Primary Examiner(s): Voorhees, Catherine
• Application Number:
  US15/439,487
• Publication Number:
  US 20170156607A1
• Time in Patent Office: 734 Days
• CPC Class Codes: A61B 5/02405 Determining heart rate variability
• CPC Class Codes: A61B 5/02411 of foetuses
• CPC Class Codes: A61B 5/0245 by using sensing means generating electric signals, i.e. ECG signals
• CPC Class Codes: A61B 5/0255 Recording instruments specially adapted therefor
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/344 Foetal cardiography
• CPC Class Codes: A61B 5/346 Analysis of electrocardiograms
• CPC Class Codes: A61B 5/35 by template matching
• CPC Class Codes: A61B 5/352 Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
• CPC Class Codes: A61B 5/4362 Assessing foetal parameters
• CPC Class Codes: A61B 5/6823 Trunk, e.g., chest, back, abdomen, hip
• CPC Class Codes: A61B 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/7221 Determining signal validity, reliability or quality A61B5/7207 and A61B5/7217 take precedence
• CPC Class Codes: A61B 5/725 using specific filters therefor, e.g. Kalman or adaptive filters A61B5/316 take precedence
• CPC Class Codes: A61B 5/7253 characterised by using transforms
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• CPC Class Codes: A61B 5/7278 Artificial waveform generation or derivation, e.g. synthesising signals from measured signals A61B5/327 takes precedence
• CPC Class Codes: A61B 7/04 Electric stethoscopes microphones, acoustic transducers therefor H04R

Abstract

The invention provides systems and methods for monitoring the wellbeing of a fetus by the non-invasive detection and analysis of fetal cardiac activity data, utilizing a plurality of acoustic sensors.

Claims

1. A computer-implemented method, comprising:

receiving, by at least one computer processor executing specific programmable instructions configured for the method, a plurality of Phonocardiogram (PCG) signals data inputs from a plurality of acoustic sensors;

transforming, by the at least one computer processor, utilizing an Independent-Component-Analysis (ICA), at least one first subset of the PCG signals data inputs to form a plurality of ICA transforms;

wavelet denoising, by the at least one computer processor, at least one second subset of the PCG signals data inputs to form a plurality of denoised PCG outputs;

transforming, by the at least one computer processor, utilizing the ICA, the plurality of denoised PCG outputs to form a plurality of denoised ICA transforms;

compiling, by the at least one computer processor, a plurality of detection heartbeat (DH) inputs, wherein the plurality of DH inputs comprises;

i) the plurality of ICA transforms, and

ii) the plurality of denoised ICA transforms;

detecting, by the at least one computer processor, locations of potential heartbeats in each of the DH inputs;

calculating, by the at least one computer processor, a plurality of confidence scores, each of the confidence scores corresponding to one of the potential heartbeats, wherein each of the confidence scores describes a probability that the corresponding one of the potential heartbeats represents an actual heartbeat and not a noise;

selecting, by the at least one computer processor, the ones of the DH inputs that include at least one actual heartbeat;

dividing, by the at least one computer processor, the selected DH inputs into at least two groups;

i) a first group of selected DH inputs containing potential fetal heartbeats, and

ii) a second group of selected DH inputs containing potential maternal heartbeats;

selecting, by the at least one computer processor, from the first group of selected DH inputs, at least one of the plurality of selected DH inputs that contains at least one actual fetal heartbeat, based on the at least one of the plurality of confidence scores corresponding to the at least one of the actual heartbeats included in the at least one of the plurality of selected DH inputs;

selecting, by the at least one computer processor, from the second group of selected DH inputs, at least one of the plurality of selected DH inputs that contains at least one actual maternal heartbeat, based on the at least one of the plurality of confidence scores corresponding to the at least one of the actual heartbeats included in the at least one of the plurality of selected DH inputs; and

outputting at least one visual indication corresponding to at least one of (a) the at least one actual maternal heartbeat and (b) the at least one actual fetal heartbeat.

2. The computer-implemented method of claim 1, wherein the wavelet denoising comprises:

deconstructing, by the at least one computer processor, each PCG signals data input to generate a plurality of transform coefficients, by iteratively passing each PCG signals data input through a succession of low and high pass filters;

identifying, by the at least one computer processor, a subset of heartbeat-carrying transform coefficients from the plurality of transform coefficients,

reconstructing, by the at least one computer processor, the subset of heartbeat-carrying transform coefficients to form the plurality of denoised PCG outputs.

3. The computer-implemented method of claim 1, wherein the detecting of the beat locations of the beats in each of the DH inputs comprises:

calculating, by the at least one computer processor, a predetermined transform filter;

iteratively repeating, by the at least one computer processor, based on a predetermined average window length of the predetermined transform filter, the steps of;

identifying, in each DH input, a subset of peaks having a predetermined shape and a predetermined size, and

selecting a group of peaks from the subset of peaks, wherein the group of peaks has the smallest variance measure;

calculating, by the at least one computer processor, initial locations of the beats.

4. The computer-implemented method of claim 1, wherein the calculating the confidence score comprises:

generating, by the at least one computer processor, a beat-by-beat heartbeat graph for each DH input based on the locations of the potential heartbeats.

5. The computer-implemented method of claim 1, wherein the dividing the plurality of selected DH inputs into the first group of selected DH inputs containing potential fetal heartbeats and the second group of selected DH inputs containing selected maternal heartbeats comprises:

clustering, by the at least one computer processor, the beats of each DH input according to each beat value; and

assigning, by the at least one computer processor, the selected DH inputs having a higher average beat rate into the first group of selected DH inputs containing fetal heartbeats.

6. A specifically programmed computer system, comprising:

at least one specialized computer machine, comprising;

a non-transient memory, electronically storing computer executable program code; and

at least one computer processor which, when executing the program code, becomes a specifically programmed computing processor that is configured to at least perform the following operations;

receiving a plurality of Phonocardiogram (PCG) signals data inputs from a plurality of acoustic sensors;

transforming, utilizing an Independent-Component-Analysis (ICA), at least one first subset of the PCG signals data inputs to form a plurality of ICA transforms;

wavelet denoising at least one second subset of the PCG signals data inputs to form a plurality of denoised PCG outputs;

transforming, utilizing the ICA, the plurality of denoised PCG outputs to form a plurality of denoised ICA transforms;

compiling a plurality of detection heartbeat (DH) inputs, wherein the plurality of DH inputs comprises;

i) the plurality of ICA transforms, and

ii) the plurality of denoised ICA transforms;

detecting locations of potential heartbeats in each of the DH inputs;

calculating a plurality of confidence scores, each of the confidence scores corresponding to one of the potential heartbeats, wherein each of the confidence scores describes a probability that the corresponding one of the potential heartbeats represents an actual heartbeat and not a noise;

selecting the ones of the DH inputs that include at least one actual heartbeat;

dividing the selected of DH inputs into at least two groups;

i) a first group of selected DH inputs containing potential fetal heartbeats, and

ii) a second group of selected DH inputs containing potential maternal heartbeats;

selecting, from the first group of selected DH inputs, at least one of the plurality of selected DH inputs that contains at least one actual fetal heartbeat, based on the at least one of the plurality of confidence scores corresponding to the at least one of the actual heartbeats included in the at least one of the plurality of selected DH inputs;

selecting, from the second group of selected DH inputs, at least one of the plurality of selected DH inputs that contains at least one actual maternal heartbeat, based on the at least one of the plurality of confidence scores corresponding to the at least one of the actual heartbeats included in the at least one particular maternal DH input of the plurality of selected DH inputs; and

outputting at least one visual indication corresponding to at least one of (a) the at least one actual maternal heartbeat and (b) the at least one actual fetal heartbeat.

7. The specifically programmed computer system of claim 6, wherein the wavelet denoising comprises:

deconstructing each PCG signals data input to generate a plurality of transform coefficients, by iteratively passing each PCG signals data input through a succession of low and high pass filters;

identifying a subset of heartbeat-carrying transform coefficients from the plurality of transform coefficients,

reconstructing the subset of heartbeat-carrying transform coefficients to form the plurality of denoised PCG outputs.

8. The specifically programmed computer system of claim 6, wherein the detecting of the beat locations of the beats in each of the DH inputs comprises:

calculating a predetermined transform filter;

iteratively repeating, based on a predetermined average window length of the predetermined transform filter, the steps of;

identifying, in each DH input, a subset of peaks having a predetermined shape and a predetermined size, and

selecting a group of peaks from the subset of peaks, wherein the group of peaks has the smallest variance measure;

calculating initial locations of the beats.

9. The specifically programmed computer system of claim 6, wherein the calculating the confidence score comprises:

generating a beat-by-beat heartbeat graph for each DH input based on the locations of the potential heartbeats.

10. The specifically programmed computer system of claim 6, wherein the dividing the plurality of selected DH inputs into the first group of selected DH inputs containing fetal heartbeats and the second group of selected DH inputs containing maternal heartbeats comprises:

clustering the beats of each DH input according to each beat value; and

assigning the selected DH inputs having a higher average beat rate into the first group of selected DH inputs containing fetal heartbeats.