Method and device for estimating morphological features of heart beats

• US 8,909,332 B2
• Filed: 01/26/2011
• Issued: 12/09/2014
• Est. Priority Date: 01/26/2010
• Status: Active Grant

• Current Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Inventors: Vitali, Andrea Lorenzo, Pessione, Marco
• Primary Examiner(s): MALAMUD, DEBORAH LESLIE
• Application Number:
  US13/014,551
• Publication Number:
  US 20110184297A1
• Time in Patent Office: 1,413 Days
• Field of Search: 600508-509, 600515-518, 600/521, 128/920, 128/923
• US Class Current: 600/521
• CPC Class Codes: A61B 5/316 Modalities, i.e. specific diagnostic methods
• CPC Class Codes: A61B 5/346 Analysis of electrocardiograms
• CPC Class Codes: A61B 5/352 Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
• CPC Class Codes: A61B 5/7207 of noise induced by motion artifacts
• CPC Class Codes: A61B 5/7264 Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems neural networks per se G06N3/00; expert systems per se G06N5/00

Abstract

A method estimates morphological features of heart beats from an ECG signal. Peaks of the R wave of the ECG are detected and classified using a parallel filtering structure. The first branch implements a bandpass filtering with cut off frequencies of about 10 Hz and 35 Hz, enhancing the signal-to-noise ratio (SNR) of the QRS complex. The second branch estimates morphological features of the heart beat from an alternating current (AC) replica of the ECG signal, that may be used to classify the beat and potentially detect arrhythmias.

Claims

1. A method, comprising:

filtering a digital replica of an electrocardiogram (ECG) signal with a passband filter with cut-off frequencies of about 10 Hz and 35 Hz, the filtering the digital replica including generating a corresponding passband signal;

filtering out a direct current component from said digital replica of the ECG signal, the filtering out the direct current component including generating a corresponding alternating current (AC) signal; and

detecting occurrence of peaks of R waves of the ECG signal by;

processing the passband signal with a finite impulse response filter having a pulse response that is linearly increasing, the processing including generating an intermediate signal; and

detecting peaks of R waves at zero-crossings between two consecutive maxima of said intermediate signal.

2. The method of claim 1, further comprising:

in response to detecting peaks of R waves of the ECG signal, generating estimations of a morphological feature of heart beats represented in the ECG signal by processing the AC signal, the morphological feature being one of a set consisting of width of QRS complex, height of peaks of R or P waves, and lengths of PR intervals.

3. The method of claim 2, wherein said estimations are generated by processing a replica of said AC signal blanked for an interval of about 50 ms after the occurrence of an R wave peak, according to the Hamilton-Tompkins algorithm.

4. The method of claim 1, wherein filtering out the direct current component from said ECG signal is performed before filtering the AC signal with the passband filter with cut-off frequencies of about 10 Hz and 35 Hz.

5. The method of claim 1, further comprising detecting disconnection from a patient of an electrode configured to detect the ECG signal, the detecting occurrence of peaks including determining that a number of samples of the ECG signal are above or below a useful signal range.

6. A non-transitory computer readable memory medium storing processor instructions configured to cause a processor to execute a method comprising:

filtering an electrocardiogram (ECG) signal with a passband filter with cut-off frequencies of about 10 Hz and 35 Hz, the filtering the ECG signal including generating a corresponding passband signal;

filtering out a direct current component from said ECG signal, the filtering out the direct current component including generating a corresponding alternating current (AC) signal;

detecting occurrence of peaks of R waves of the ECG signal by processing the passband signal; and

in response to detecting peaks of R waves of the ECG signal, generating estimations of a morphological feature of heart beats represented in the ECG signal by processing a replica of said AC signal blanked for an interval of about 50 ms after the occurrence of an R wave peak, according to the Hamilton-Tompkins algorithm, the morphological feature being one of a set consisting of width of QRS complex, height of peaks of R or P waves, and lengths of PR intervals.

7. The non-transitory computer readable memory medium of claim 6, wherein the filtering steps are performed on a digital replica of the ECG signal, and said peaks of R waves of the ECG signal are detected by:

processing said passband signal with a slope filter, generating an intermediate signal; and

detecting peaks of R waves at zero-crossings between two consecutive maxima of said intermediate signal.

8. The non-transitory computer readable memory medium of claim 6, wherein the method further comprises detecting disconnection from a patient of an electrode configured to detect the ECG signal, the detecting disconnection including determining that a number of samples of the ECG signal are above or below a useful signal range.

9. An ECG signal monitor, comprising:

a passband filter, with cut-off frequencies of about 10 Hz and 35 Hz, configured to filter an electrocardiogram (ECG) signal and generate a corresponding passband signal;

a DC filter configured to filter out a direct current component from said ECG signal and generate a corresponding alternating current (AC) signal;

a peak detector configured to detect occurrence of peaks of R waves of the ECG signal by processing the passband signal; and

a morphological feature extractor configured to, in response to the peak detector detecting peaks of R waves of the ECG signal, generate estimations of a morphological feature of heart beats represented in the ECG by processing a replica of said AC signal blanked for an interval of about 50 ms after the occurrence of an R wave peak, according to the Hamilton-Tompkins algorithm, the morphological feature being one of a set consisting of width of QRS complex, height of peaks of R or P waves, and lengths of PR intervals.

10. The monitor of claim 9, wherein the peak detector includes:

a slope filter configured to detect slope of the passband signal and generate an intermediate signal; and

a moving average filter configured to process said intermediate signal and generate a moving average of a number of samples of the intermediate signal.

11. The monitor of claim 9, further comprising:

a plurality of electrodes configured to detect an analog ECG signal of a patient;

an analog circuit configured to process the analog ECG signal; and

an analog/digital converter configured to convert the analog ECG signal to a digital ECG signal, wherein the DC filter is configured to process the digital ECG signal.

12. The monitor of claim 11, further comprising:

an electrode disconnect detector configured to detect disconnection of one or more of the electrodes from the patient by determining that a number of samples of the digital ECG signal are above or below a useful signal range.