System and method for automated identification of abnormal ciliary motion

Patent metadata

Field	Value
Publication number	US-9483685-B2
Application number	US-201514697742-A
Country	US
Kind code	B2
Filing date	Apr 28, 2015
Priority date	Apr 28, 2014
Publication date	Nov 1, 2016
Grant date	Nov 1, 2016

Abstract

A method of classifying ciliary motion includes receiving digital video data representing the ciliary motion generated by an image capture device, wherein the digital video data includes a plurality of frames. The method further includes receiving an indication of a region of interest applicable to each of the frames, wherein the region of interest includes a plurality of pixels in each of the frames, calculating time series elemental motion data for at least one elemental motion parameter for the region of interest based on the digital video data, and using the time series elemental motion data to classify the ciliary motion.

First claim

What is claimed is:

1. A method of classifying ciliary motion, comprising:
receiving digital video data representing the ciliary motion generated by an image capture device, the digital video data comprising a plurality of frames;
receiving an indication of a region of interest applicable to each of the frames, the region of interest comprising a plurality of pixels in each of the frames;
calculating time series elemental motion data for at least one elemental motion parameter for the region of interest based on the digital video data; and
using the time series elemental motion data to classify the ciliary motion.

2. The method according to claim 1 , wherein the calculating the time series elemental motion comprises creating optical flow vector data for the region of interest using the digital video data and calculating the time series elemental motion data for the at least one elemental motion parameter based on the optical flow vector data.

3. The method according to claim 2 , wherein the using the time series elemental motion data comprises creating at least one temporal domain histogram and at least one frequency domain histogram using the times series elemental motion data, and using the at least one temporal domain histogram and the at least one frequency domain histogram to classify the ciliary motion.

4. The method according to claim 2 , wherein the calculating the time series elemental motion data for at least one elemental motion parameter based on the optical flow vector data comprises calculating time series rotation data and time series deformation data based on the optical flow vector data.

5. The method according to claim 4 , wherein the using step comprises creating a rotation temporal domain histogram and a rotation frequency domain histogram using the time series rotation data and a deformation temporal domain histogram and a deformation frequency domain histogram using the time series deformation data, and using the rotation temporal domain histogram, the rotation frequency domain histogram, the deformation temporal domain histogram, the deformation frequency domain histogram to classify the ciliary motion.

6. The method according to claim 1 , wherein the using the time series elemental motion data to classify the ciliary motion comprises calculating a plurality of autoregressive motion parameters using the time series elemental motion data and using the autoregressive motion parameters to classify the ciliary motion.

7. The method according to claim 6 , wherein the calculating the plurality of autoregressive motion parameters comprises performing singular value decomposition using the time series elemental motion data to generate a plurality of basis vectors, and generating the autoregressive motion parameters using the basis vectors.

8. The method according to claim 1 , wherein the using the time series elemental motion data to classify the ciliary motion comprises creating at least one temporal domain histogram and at least one frequency domain histogram using the time series elemental motion data, calculating a plurality of autoregressive motion parameters using the time series elemental motion data, and using the at least one temporal domain histogram, the at least one frequency domain histogram and the autoregressive motion parameters to classify the ciliary motion.

9. The method according to claim 1 , wherein the ciliary motion is classified as normal or abnormal in the using step.

10. The Method according to claim 1 , further comprising generating and displaying an output based on a result of the using step.

11. The method according to claim 1 , further comprising identifying a second region of interest applicable to each of the frames, the second region of interest comprising a second plurality of pixels in each of the frames, and calculating second times elemental motion data for at least one elemental motion parameter for the second region of interest based on the digital video data, wherein the using step comprises using the elemental motion data and the second elemental motion to classify the ciliary motion.

12. A non-transitory computer readable medium storing one or more programs, including instructions, which when executed by a computer, causes the computer to perform the method of claim 1 .

13. An apparatus for classifying ciliary motion, comprising:
an image capture device for generating digital video data representing the ciliary motion, the digital video data comprising a plurality of frames;
a processor apparatus that stores and is structured to execute a number of routines, the number of routines being structured to:
receive the digital video data;
receive an indication of a region of interest applicable to each of the frames, the region of interest comprising, a plurality of pixels in each of the frames;
calculate time series elemental motion data for at least one elemental motion parameter for the region of interest based on the digital video data; and
use the time series elemental motion data to classify the ciliary motion.

14. The apparatus according to claim 13 , wherein the routines are structured to calculate the time series elemental motion by creating optical flow vector data for the region of interest using the digital video data and calculating the time series elemental motion data for the at least one elemental motion parameter based on the optical flow vector data.

15. The apparatus according to claim 14 , wherein the routines are structured to use the time series elemental motion data by creating at least one temporal domain histogram and at least one frequency domain histogram using the times series elemental motion data, and using the at least one temporal domain histogram and the at least one frequency domain histogram to classify the ciliary motion.

16. The apparatus according to claim 14 , wherein the routines are structured to calculate the time series elemental motion data for at least one elemental motion parameter based on the optical flow vector data comprises calculating time series rotation data and time series deformation data based on the optical flow vector data.

17. The apparatus according to claim 16 , wherein the routines are structured to use the time series elemental motion data by creating a rotation temporal domain histogram and a rotation frequency domain histogram using the time series rotation data and a deformation temporal domain histogram and a deformation frequency domain histogram using the time series deformation data, and using the rotation temporal domain histogram, the rotation frequency domain histogram, the deformation temporal domain histogram, the deformation frequency domain histogram to classify the ciliary motion.

18. The apparatus according to claim 13 , wherein the routines are structured to use the time series elemental motion data to classify the ciliary motion by calculating a plurality of autoregressive motion parameters using the time series elemental motion data and using the autoregressive motion parameters to classify the ciliary motion.

19. The apparatus according, to claim 18 , wherein the routines are structured to calculate the plurality of autoregressive motion parameters by performing singular value decomposition using the time series elemental motion data to generate a plurality of basis vectors, and generating the autoregressive motion parameters using the basis vectors.

20. The apparatus according to claim 13 , wherein the routines are structured to use the time series elemental motion data to classify the ci

Assignees

• Univ Of Pittsburgh—Of The Commonwealth System Of Higher Education
• Univ Of Pittsburgh - Of The Commonwealth System Of Higher Education

Inventors

• Chennubhotla Chakra S
• Quinn Shannon P
• Durkin John R
• Zahid Maliha
• Francis Richard John Bruce
• Ya Lo Cecilia Wen

Classifications

• G06T7/0016 Primary

  involving temporal comparison · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V20/698

  Matching; Classification · CPC title

• G06F18/2411

  based on the proximity to a decision surface, e.g. support vector machines · CPC title

• G06V10/431

  Frequency domain transformation; Autocorrelation · CPC title