Medical decision support system
Inventors
TELENKOV, Sergey A. • Castelino, Robin F. • GLOAG, David • LABONTÉ, Daniel • ISLAM, Md Shahidul
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11896380-B2
Publication Date
2024-02-13
Expiration Date
Abstract
An auscultatory sound signal from at least one auscultatory sound-or-vibration sensor is filtered with a high-pass filter and then segmented into a plurality of associated heart cycle segments responsive to associated R-peak locations of an electrographic envelope signal representing an envelope response to an even power of an associated electrographic signal from an ECG sensor. A representation an envelope responsive to an even power of said auscultatory sound signal within said at least one heart cycle is locally modeled about at least a second peak to provide for locating the start of diastole of said at least one heart cycle.
Core Innovation
The invention concerns segmenting an auscultatory sound signal using an ECG-derived timing reference. An electrographic signal is received from an ECG sensor, and an electrographic envelope signal is generated that represents an envelope responsive to an even power of the electrographic signal. Peaks of the electrographic envelope are located corresponding to R-peaks, so that heart-cycle segmentation is responsive to ECG-derived peaks.
At least one auscultatory sound signal is received from a corresponding auscultatory sound-or-vibration sensor and filtered with a high-pass filter to generate at least one high-pass-filtered auscultatory sound signal. The high-pass-filtered auscultatory sound signal is segmented into at least one heart-cycle segment responsive to the plurality of peaks of the electrographic envelope signal. Within each heart-cycle segment, an auscultatory envelope signal is generated as an envelope responsive to an even power of the auscultatory sound signal, and at least a second peak corresponding to a second heart sound is located.
A local mathematical model of at least the second peak of the auscultatory envelope signal is generated, wherein the local mathematical model comprises a quadratic equation. At least one root selected from a first root and a second root of the local mathematical model is located to provide for determining an associated start of diastole of the heart-cycle segment. This produces diastole timing boundaries from localized modeling around the second heart-sound peak.
Claims Coverage
Two independent claims are present, both covering a segmentation method that aligns auscultatory signal heart-cycle segments to ECG-derived R-peaks via an electrographic envelope based on an even power, and then determines a start of diastole from a second heart-sound peak using a local quadratic model and selected roots. A shared set of inventive features includes ECG envelope peak locating, auscultatory high-pass filtering, within-beat auscultatory envelope peak locating, and diastole boundary determination by quadratic-root selection; dependent claims add refinements such as peak validation and diastole noise/quality evaluation.
ECG envelope based on even power for R-peak correspondence
Generating an electrographic envelope signal representing an envelope responsive to an even power of an electrographic signal and locating peaks of the electrographic envelope signal corresponding to R-peaks of the electrographic signal.
High-pass filtering of auscultatory signal for beat segmentation
Filtering at least one auscultatory sound signal with a high-pass filter to generate a corresponding at least one high-pass-filtered auscultatory sound signal, and segmenting the high-pass-filtered auscultatory sound signal into at least one heart-cycle segment responsive to the plurality of peaks of the electrographic envelope signal.
Auscultatory envelope based on even power with second heart-sound peak detection
Within each heart-cycle segment, generating an auscultatory envelope signal representing an envelope responsive to an even power of the auscultatory sound signal and locating at least a second peak of the auscultatory envelope signal corresponding to a second heart sound.
Quadratic local mathematical model and root selection for start of diastole
Generating a local mathematical model of at least the second peak of the auscultatory envelope signal, wherein the local mathematical model comprises a quadratic equation, and locating at least one root selected from a first root and a second root of the local mathematical model to provide for determining an associated start of diastole of the heart-cycle segment.
Constrained quadratic model point selection around a peak
Generating a local mathematical model wherein the quadratic equation is formed using a set of points selected around a chosen peak such that the magnitudes of the selected points are not in excess of a specified fraction of the peak magnitude.
Threshold-based second peak selection and adjustment
Locating the at least a second peak by comparing the auscultatory envelope signal to an adjustable magnitude threshold, selecting peak values exceeding the threshold, and iteratively adjusting the threshold so that only the corresponding peak remains above the threshold for the heart-cycle segment.
Time-duration validation between first and second heart sound peaks
Locating at least a second peak in the auscultatory envelope signal, locating a first peak corresponding to a first heart sound, and validating the peaks by comparing a time duration between the first and second peaks to a threshold.
Diastole region noise/quality check using mean power and noise thresholds
Defining a diastole region for at least one heart-cycle segment and analyzing the diastole region by determining whether a measure based on mean power exceeds one or more noise thresholds.
Low-pass filtering frequency constraint at 500 Hz
Using a low-pass filter configured to remove significant acoustic energy above 500 Hz.
Across both independent claims, the inventive thrust is ECG-driven heart-cycle segmentation using an even-power electrographic envelope to locate R-peaks, followed by per-beat auscultatory envelope processing to locate a second heart-sound peak and a local quadratic model around that peak. Selecting a root of the quadratic model determines an associated start of diastole, with dependent claims further refining peak selection/validation and adding diastole-region noise/quality evaluation and frequency-filter constraints.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Not explicitly described in patent.
Interested in licensing this patent?