Method and system for signal analysis

Patent metadata

Field	Value
Publication number	US-9659229-B2
Application number	US-201414767507-A
Country	US
Kind code	B2
Filing date	Jan 28, 2014
Priority date	Feb 12, 2013
Publication date	May 23, 2017
Grant date	May 23, 2017

Abstract

An image of a human, animal or machine subject, is analysed to detect regions which include strong periodic intensity variations, such as a photoplethysmogram (PPG) signal in a human or animal, or some periodic vibration in a machine. The image is divided into plural regions of fixed order is fitted to a representative intensity signal for that region. The poles of the fitted autoregressive model are thresholded by magnitude to select only the pole or poles with a magnitude greater than the threshold. The pole magnitude therefore acts as a signal quality index. The dominant pole is representative of the strongest periodic information and the frequency of that spectral component can be derived from the phase angle of the pole. The image may be redisplayed with image attributes, e.g. color-coding, according to the pole magnitude in each region of interest and/or the dominant pole phase angle in each region of interest. In the case of a PPG image signal this can give maps of heart rate and breathing rate.

First claim

The invention claimed is:

1. A method of analysing an image of a subject to identify spatial areas which contain periodic intensity variations representing valid temporal information about the subject, comprising the steps of:
dividing said image into a plurality of spatial areas and obtaining respective signals representative of an image intensity for each spatial area;
spectrally analysing the intensity variations of each representative signal by fitting respective autoregressive models to the representative signals, each of the fitted autoregressive models comprising a plurality of poles representing spectral components of the intensity variations of each representative signal, each pole having a magnitude dependent upon the strength of the spectral component and a phase angle dependent upon the frequency of the spectral component; and
selecting those spatial areas whose fitted autoregressive model has a dominant pole whose magnitude is greater than a predetermined threshold and identifying those spatial areas as containing periodic intensity variations representing said valid temporal information.

2. A method according to claim 1 wherein the spatial areas are m by n pixels of the image where m and n are positive integers.

3. A method according to claim 1 wherein the representative signal is an average or modal intensity over the spatial area.

4. A method according to claim 1 wherein the image is a red-green-blue colour image and intensity variations in at least one of the three components are spectrally analysed.

5. A method according to claim 1 further comprising the step of identifying and ignoring poles representing spectral components outside an expected frequency range for valid temporal information.

6. A method according to claim 1 further comprising the step of identifying and ignoring poles representing ambient light intensity variations due to aliasing.

7. A method according to claim 1 further comprising the step of selecting the dominant pole as being either the pole of highest magnitude or the pole of smallest phase angle whose magnitude satisfies the threshold, and calculating from the phase angle the frequency of the intensity variation represented by the selected pole.

8. A method according to claim 7 further comprising identifying peaks in the frequency spectrum of the intensity variations and selecting as the dominant pole the pole which corresponds to the highest magnitude peak.

9. A method according to claim 1 further comprising defining a time window in each of said representative signals, performing the spectral analysis and outputting the temporal information identified as valid, advancing the window and repeating the spectral analysis and outputting the temporal information identified as valid.

10. A method according to claim 1 further comprising displaying the image with a display attribute of the spatial areas being set to represent visually the magnitude of a selected pole in the fitted autoregressive model for that spatial area.

11. A method according to claim 10 wherein:
the display attribute defines the displayed colour of each one of the spatial areas, and
the displayed colour of each one of the spatial areas is based on the magnitude of the selected pole in each one of the spatial areas.

12. A method according to claim 1 further comprising displaying the image with a display attribute of the spatial areas being set to represent visually the frequency of a pole of the fitted autoregressive model for that spatial area whose magnitude is greater than the predetermined threshold.

13. A method according to claim 12 wherein:
the display attribute defines the displayed colour of each one of the spatial areas, and
the displayed colour of each one of the spatial areas is based on the frequency of the dominant pole in each one of the spatial areas.

14. A method according to claim 1 wherein the subject is human or animal and the image contains a photoplethysmogram image.

15. A method according to claim 14 wherein the temporal information is at least one of heart rate and breathing rate.

16. A method according to claim 1 further comprising segmenting the image on the basis of at least one of the magnitude and phase angle of at least one of the poles in each of said spatial areas.

17. A non-transitory computer-readable medium comprising program code to analyse an image of a subject to identify spatial areas that contain periodic intensity variations representing valid temporal information about the subject, including:
dividing the image into a plurality of spatial areas and obtaining respective signals representative of an image intensity for each spatial area;
spectrally analysing the intensity variations of each representative signal by fitting respective autoregressive models to the representative signals, each of the fitted autoregressive models comprising a plurality of poles representing spectral components of the intensity variations of each representative signal, each pole having a magnitude dependent upon the strength of the spectral component and a phase angle dependent upon the frequency of the spectral component; and
selecting those spatial areas whose fitted autoregressive model has a dominant pole whose magnitude is greater than a predetermined threshold and identifying those spatial areas as containing periodic intensity variations representing said valid temporal information.

18. An image analysis system for analysing an image of a subject to identify spatial areas which contain periodic intensity variations representing valid temporal information about the subject, comprising:
a processor and associated memory,
wherein the memory stores computer instructions that, when executed by the processor, cause the processor to perform operations including:
dividing an image into a plurality of spatial areas and obtaining respective signals representative of an image intensity for each spatial area;
spectrally analysing the intensity variations of each resresentative signal by fitting respective autoregressive models to the representative signals, each of the fitted autoregressive models comprising a plurality of poles representing spectral components of the intensity variations of each representative signal, each pole having a magnitude dependent upon the strength of the spectral component and a phase angle dependent upon the frequency of the spectral component; and
selecting those spatial areas whose fitted autoregressive model has a dominant pole whose magnitude is greater than a predetermined threshold and identifying those spatial areas as containing periodic intensity variations representing said valid temporal information.

19. A vital sign monitor comprising an image analysis system according to claim 18 .

Assignees

• Isis Innovation

Inventors

• Clifton David Andrew
• Montoya Mauricio Christian Villarroel
• Tarassenko Lionel

Classifications

• G06T2207/30201

  Face · CPC title

• G06T2207/30104

  Vascular flow; Blood flow; Perfusion · CPC title

• G06T2207/30088

  Skin; Dermal · CPC title

• G06T2207/30076

  Plethysmography · CPC title

• G06T2207/20021

  Dividing image into blocks, subimages or windows · CPC title