Cardiac diagnostic system and method

• US 20050222508A1
• Filed: 04/05/2004
• Published: 10/06/2005
• Est. Priority Date: 04/05/2004
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Moreno, Pedro J., Goddeau, David, Logan, Beth T.
• Granted Patent: US 7,174,205 B2
• Time in Patent Office: Days
• US Class Current: 600/509
• CPC Class Codes: A61B 5/35 by template matching
• 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
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

Computer method and apparatus of analyzing ECG signals of a subject include receiving a subject electrocardiogram signal and comparing it against signal patterns of known cardiac syndromes. A library of example predefined signals is employed. Distance measures indicating similarity of the subject signal to the example predefined signals are produced and form a sequence of vectors. The sequence of vectors are input into a classifier which determines existence of signal patterns indicative of any cardiac syndromes in the subject.

Claims

1. A method of analyzing ECG signals of a subject comprising the computer implemented steps of:

receiving a subject electrocardiogram signal to be analyzed;

using a library of example predefined signals indicative of known cardiac syndromes, producing distance measures indicating similarity of the subject signal to the example predefined signals;

forming a sequence of vectors from the produced distance measures; and

using the sequence of vectors as input into a classifier which determines therefrom existence of any cardiac syndromes in the subject.

2. A method of analyzing as claimed in claim 1, further comprising the step of segmenting the subject signal into working segments;

wherein the step of producing distance measures includes for each working segment, forming a vector with a same number of components as number of example signals in the library, each component corresponding to a different one of the example signals and having a respective distance value indicative of how similar the working segment is to the respective example signal.

3. A method of analyzing as claimed in claim 1, wherein the step of producing distance measures includes:

segmenting the subject signal into a plurality of segments; and

for each segment, (a) computing distance between data points of the segment and data points of each example signal of the library such that a vector of distances is formed for each example signal, and (b) defining a distance measure for the segment as a function of distance between the formed vectors of distances.

4. A method of analyzing as claimed in claim 1, further comprising the step of:

labeling segments of an example signal in the library based on output of the classifier.

5. A method of analyzing as claimed in claim 1, wherein the library of example signals holds samples defined by cardiologists, each sample having a respective cardio syndrome and a corresponding example cardiogram signal.

6. Computer apparatus for analyzing ECG signals of a subject comprising:

a module for receiving a subject electrocardiogram signal to be analyzed;

a library of example predefined signals;

a kernel function member coupled between the library and the receiving module, the kernel function member (i) producing distance measures indicating similarity of the subject signal to the example predefined signals, and (ii) forming a sequence of vectors from the produced distance measures; and

a classifier using the formed sequence of vectors as input and determining existence of any cardiac syndromes in the subject.

7. Computer apparatus as claimed in claim 6 wherein the receiving module segments the subject signal into working segments;

and

the kernel function member produces a distance measure for each working segment by forming a vector with a same number of components as number of example signals in the library, each component corresponding to a different one of the example signals and having a respective distance value indicative of how similar the working segment is to the respective example signal.

8. Computer apparatus as claimed in claim 6 wherein the receiving module segments the subject signal into working segments;

and

the kernel function member for each working segment, (a) computes distance between data points of the segment and data points of each example signal of the library such that a vector of distances is formed for each example signal, and (b) defines a distance measure for the working segment as a function of distance between the formed vectors of distances.

9. Computer apparatus as claimed in claim 6 wherein segments of the example signals in the library are labeled based on output of the classifier.

10. Computer apparatus as claimed in claim 6, wherein the library of example signals holds samples defined by cardiologists, each sample having a respective cardiac syndrome and a corresponding example cardiogram signal.

11. Computer apparatus for analyzing ECG signals of a subject comprising:

means for receiving a subject ECG signal;

library means for storing example predefined signals indicative of known cardiac syndromes;

kernel function means responsive to the subject ECG signal for (i) producing distance measures indicating similarity of the subject signal to the example predefined signals and (ii) forming a sequence of vectors from the produced distance measures; and

classifying means for using the formed sequence of vectors as input and determining existence of any cardiac syndromes in the subject.

12. Computer apparatus as claimed in claim 11 wherein the means for receiving includes means for segmenting the subject signal into working segments.

13. Computer apparatus as claimed in claim 12 wherein the kernel function means produces a distance measure for each working segment by forming a vector with a same number of components as number of example signals in the library, each component corresponding to a different one of the example signals and having a respective distance value indicative of how similar the working segment is to the respective example signal.

14. Computer apparatus as claimed in claim 12 wherein the kernel function means for each working segment (a) computes distance between data points of the segment and data points of each example signal of the library such that a vector of distances is formed for each example signal, and (b) defines a distance measure for the working segment as a function of distance between the formed vectors of distances.

15. Computer apparatus as claimed in claim 11 wherein segments of the example signals in the library are labeled based on output of the classifying means.

16. Computer apparatus as claimed in claim 11 wherein the library means holds examples defined by cardiologists, each example having a respective cardiac syndrome and a corresponding example cardiogram signal.

17. A system for analyzing ECG signals of a subject comprising:

library means for providing example predefined signals indicative of known cardiac syndromes;

measuring means for producing distance measures indicating similarity of a subject signal to the example predefined signals;

vector means for forming a sequence of vectors from the produced distance for use of the sequence of vectors as input into a classifier which determines therefrom existence of any cardiac syndromes in the subject.

18. A system as claimed in claim 17 further comprising segmenting means for segmenting the subject signal into working segments;

wherein the measuring means for each working segment forms a vector with a same number of components as number of example signals in the library means, each component corresponding to a different one of the example signals and having a respective distance value indicative of how similar the working segment is to the respective example signal.

19. A system as claimed in claim 17 wherein the measuring means segments the subject signal into a plurality of segments;

and

for each segment, (a) computes distance between data points of the segment and data points of each example signal of the library means such that a vector of distances is formed for each example signal, and (b) defines a distance measure for the segment as a function of distance between the formed vectors of distances.

20. A system as claimed in claim 17 further comprising labeling means for labeling segments of an example signal in the library means based on output of the classifier.