Method and apparatus for waveform assessment

• US 7,043,293 B1
• Filed: 12/24/2002
• Issued: 05/09/2006
• Est. Priority Date: 12/24/2002
• Status: Active Grant

• Current Assignee: FUJIFILM SonoSite, Inc. (Fujifilm Holdings Corporation)
• Original Assignee: Cardiodynamics International Corporation (Fujifilm Holdings Corporation)
• Inventors: Baura, Gail D.
• Primary Examiner(s): Getzow, Scott M.
• Application Number:
  US10/329,129
• Time in Patent Office: 1,232 Days
• Field of Search: 600508-521
• US Class Current: 600/509
• 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 5/7203 for noise prevention, reduction or removal
• CPC Class Codes: A61B 5/726 using Wavelet transforms
• 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: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

An improved method and apparatus for assessing time-variant waveforms and identifying artifacts of interest therein. In one exemplary embodiment, an iterative interval search technique is applied to ECG waveform data in order to identify one or more artifacts (e.g., right atrial and right ventricular “spikes”) within the waveform in conjunction with a fuzzy logic noise threshold analysis. This technique allows for robust artifact identification in waveforms where significant variations in noise and artifact periodicity may exist. The identified artifact(s) may then be used as an input to another process, such as being substituted as an ECG “Q” point for subsequent fiducial point detection within a cardiographic impedance waveform. Apparatus including computer programs for implementing the aforementioned techniques are also disclosed.

Claims

1. A method of identifying an artifact in a cardiographic waveform, comprising:

providing at least one cardiographic waveform;

determining the shape of at least one of said artifacts within said waveform using an iterative interval search; and

identifying said artifact based at least in part on said shape;

wherein said act of identifying further comprises analyzing said shape with respect to a variable noise threshold.

2. A method of identifying an artifact in a cardiographic waveform, comprising:

providing at least one cardiographic waveform;

determining the shape said artifacts within said waveform; and

identifying said artifact based at least in part on said shape;

wherein said act of identifying further comprises analyzing the noise threshold associated with said at least one waveform using fuzzy logic.

3. A method of identifying at least one artifact within a waveform, comprising:

sampling said waveform at a first frequency to produce sampled data;

filtering said sampled data according to a first pass band to produce first filtered data;

filtering said first filtered data according to a second pass band to produce second filtered data;

determining at least one noise threshold based at least in part on said second filtered data; and

identifying at least one artifact from said waveform based at least in part on said at least one noise threshold.

4. The method of claim 3, wherein said act of identifying at least one artifact comprises performing at least one golden section search.

5. The method of claim 3, wherein said act of determining at least one noise threshold comprises analyzing at least a portion of said second filtered using fuzzy logic.

6. A method of processing waveform data, comprising:

sampling said waveform at a first frequency to produce sampled data;

filtering said sampled data according to at least one filter scheme to produce filtered data; and

determining at least one noise threshold associated with said waveform based at least in part on fuzzy logic analysis of said filtered data.

7. The method of claim 6, wherein said act of filtering comprises:

high pass filtering said sampled data according to a first pass band; and

low pass filtering said high pass filtered data according to a second pass band.

8. The method of claim 6, wherein said act of determining at least one noise threshold comprises:

fuzzification of at least a portion of said filtered data to produce a plurality of fuzzy inputs;

mapping at least a portion of said fuzzy inputs to a plurality of fuzzy outputs; and

determining said at least one noise threshold based at least on part on said plurality of fuzzy outputs.

9. The method of claim 8, wherein said act of fuzzification comprises generating said plurality of fuzzy inputs based on at least one membership function and said at least portion of filtered data.

10. The method of claim 8, wherein said act of mapping comprises mapping using a rule based inference (RBI) process.

11. The method of 8, wherein said act of determining said at least one noise threshold comprises:

providing at least one output membership function; and

defuzzification of said plurality of fuzzy outputs into at least one crisp noise threshold value.

12. Apparatus adapted for analysis of physiologic data, comprising:

at least one data input, said at least one input adapted for operative coupling to at least one device adapted to produce signals derived from a living subject;

a data processor operatively coupled to receive data from said at least one input; and

a computer program adapted to run on said at least one processor, said computer program further comprising;

a filtering algorithm;

a noise threshold algorithm; and

an artifact detector algorithm.

13. The apparatus of claim 12, wherein said computer program further comprises:

a delay adapter algorithm, and

an artifact discriminator algorithm.

14. The apparatus of claim 12, wherein said computer program further comprises a pace algorithm adapted for associating at least one detected artifact within a waveform with another event within said waveform.

15. The apparatus of claim 12, wherein said noise threshold algorithm comprises at least one fuzzy logic analysis algorithm.

16. The apparatus of claim 15, wherein said artifact detector algorithm comprises at least one golden section search algorithm.

17. The apparatus of claim 12, wherein said artifact detector algorithm comprises at least one golden section search algorithm.

18. The apparatus of claim 12, wherein said at least one device adapted to produce signals comprises a plurality of electrodes, each of said electrodes having a plurality of terminals spaced at a predetermined distance.

19. The apparatus of claim 12, wherein said apparatus comprises an impedance cardiography (ICG) device, and said data comprises both ECG and impedance data.

20. Apparatus adapted to identify an artifact in physiologic waveforms from a living subject, comprising:

a digital processor adapted to run algorithms thereon;

at least one electrode in operative communication with said processor and adapted to provide at least one physiologic waveform from said subject;

a fuzzy logic algorithm adapted to identify at least one noise threshold within said at least one physiologic waveform; and

an iterative interval search algorithm adapted to identify said artifact based at least in part on said noise threshold and said waveform.

21. A method of dynamically adjusting the delay associated with a plurality of waveform artifacts, comprising:

providing a first delay value;

evaluating said first delay value with respect to at least two of said waveform artifacts during a first iteration; and

adjusting said first delay value to a second delay value during a second iteration based at least in part on said act of evaluating.

22. The method of claim 21, wherein said act of evaluating comprises determining the number of occurrences of said artifacts with a pre-selected sampling interval.

23. The method of claim 22, wherein said method comprises dynamically adjusting the delay associated with the atrial and ventricular spikes associated with an ECG waveform.

24. A method of discriminating between a plurality of artifacts in a waveform, comprising:

identifying a plurality of artifacts;

classifying individual ones of said artifacts according to a first classification having at least two types; and

evaluating said artifacts based on said first classification by;

if one or more artifacts of only one of said types is present, assigning at least one of said artifacts of said one type as an artifact of interest;

or

if at least one artifact of each of said at least two types are present, assigning an artifact of a preselected one of said at least two types as an artifact of interest.

25. The method of claim 24, wherein said act of classifying according to at least two types comprises classifying said artifacts as either atrial or ventricular spikes.

26. The method of claim 24, wherein said act of assigning an artifact of a preselected type comprises assigning a ventricular spike as the Q point in an ECG waveform.

27. A method of discriminating between a plurality of artifacts in an electrocardiographic waveform, comprising:

identifying a plurality of artifacts;

classifying, using at least a computer program, individual ones of said artifacts according to a first classification having multiple types; and

evaluating, using at least a computer program, said artifacts based on said first classification by;

if one or more artifacts of only one of said multiple types is present, assigning at least one of said artifacts of said one type as an artifact of interest;

or

if at least one artifact of each of said multiple types are present, assigning at least one artifact of a preselected one of said multiple types as an artifact of interest.

28. Apparatus adapted to identify an artifact in physiologic waveforms from a living subject, comprising:

digital processor means for running algorithms thereon;

at least one sensor means, in operative communication with said processor means, for providing at least one physiologic waveform from said subject;

fuzzy logic algorithm means adapted to identify at least one noise threshold within said at least one physiologic waveform; and

iterative interval search algorithm means adapted to identify said artifact based at least in part on said noise threshold and said waveform.

29. A method of detecting an artifact in a waveform, said artifact having at least one feature, comprising:

providing at least one waveform;

determining at least one noise threshold with respect to said at least one waveform; and

detecting said artifact based at least in part on said at least one noise threshold and said at least one feature, said detecting comprising performing at least one iterative interval search.

30. The method of claim 29, wherein said act of performing at least one iterative interval search comprises identifying the shape of said artifact using said search.

31. The method of claim 30, wherein said at least one iterative interval search comprises iteratively evaluating a search interval using a golden section search, and said act of identifying the shape comprises identifying at least two points at which function values are compared, said at least two points being symmetrically located in said search interval during each of said search iterations.

32. The method of claim 30, further comprising specifying at least one uncertainty interval.

33. The method of claim 32, further comprising reducing the length of said at least one uncertainty interval by given value.

34. The method of claim 33, wherein said given value comprises a fixed ratio.

35. A method of detecting an artifact in a ECG waveform based on shape, comprising:

providing at least one ECG waveform;

determining a noise threshold value using at least one fuzzy variable; and

detecting said artifact based at least in part on said shape and said at least one noise parameter.

36. A method of detecting an artifact in a physiologic waveform based on shape, comprising:

providing at least one physiologic waveform;

determining at least one noise parameter with respect to said at least one waveform; and

detecting said artifact based at least in part on said shape and said at least one noise parameter, said act of detecting comprising performing at least iterative interval search.

37. Apparatus adapted to identify an artifact in physiologic waveforms from a living subject, comprising:

a digital processor adapted to run computer programs thereon;

at least one electrode in operative communication with said processor and adapted to provide at least one physiologic waveform from said subject; and

a computer program adapted to;

determine the shape of said artifact within said at least one waveform; and

identify said artifact based at least in part on said shape; and

selectively substitute said artifact for at least one fiducial point within said at least one waveform.

38. The apparatus of claim 37, wherein said computer program is further adapted to determine a noise threshold associated with said at least one waveform.

39. The apparatus of claim 37, wherein said at least one fiducial point comprises the Q point within an ECG waveform.

40. The apparatus of claim 39, wherein said computer program is further adapted to utilize said substituted artifact in the definition of a search interval used with another waveform.

41. A method of detecting an artifact in a waveform, said artifact having at least one feature, comprising:

a step for providing at least one waveform;

a step for determining at least one noise threshold with respect to said at least one waveform; and

a step for detecting said artifact based at least in part on said at least one noise threshold and said at least one feature, said set for detecting further comprising performing at least one iterative interval search including at least identifying the shape of said artifact using said search, said at least one iterative interval search comprising iteratively evaluating a search interval using a golden section search, said act of identifying the shape comprising identifying at least two points at which function values are compared, said at least two points being symmetrically located in said search interval during each of a plurality of search iterations;

wherein said step for providing at least one waveform comprises a step for providing an ECG, and said step for detecting an artifact comprises a step for detecting a feature related to the coronary function of a living subject.

42. An apparatus for discriminating between a plurality of artifacts in an electrocardiographic waveform, comprising:

means for identifying a plurality of artifacts;

means for classifying, using at least a computer program, individual ones of said artifacts according to a first classification having multiple types; and

means for evaluating, using at least a computer program, said artifacts based on said first classification by;

if one or more artifacts of only one of said multiple types is present, assigning at least one of said artifacts of said one type as an artifact of interest;

or

if at least one artifact of each of said multiple types are present, assigning at least one artifact of a preselected one of said multiple types as an artifact of interest.

43. An apparatus for detecting an artifact in a waveform, said artifact having at least one feature, comprising:

means for providing at least one waveform;

means for determining at least one noise threshold with respect to said at least one waveform; and

means for detecting said artifact based at least in part on said at least one noise threshold and said at least one feature, said means for detecting being adapted to perform at least one iterative interval search.

44. The apparatus of claim 43, wherein said means for performing at least one iterative interval search comprises:

means for identifying the shape of said artifact based at least in part on said search.

45. The apparatus of claim 44, wherein said means for performing at least one iterative interval search comprises:

means for iteratively evaluating a search interval using a golden section search, and said act of identifying the shape comprises identifying at least two points at which function values are compared, said at least two points being symmetrically located in said search interval during each of said search iterations.

46. The apparatus of claim 45, further comprising specifying at least one uncertainty interval.

47. The apparatus of claim 46, further comprising reducing the length of said at least one uncertainty interval by given value.

48. The apparatus of claim 47, wherein said given value comprises a fixed ratio.