System and method for the selection of optical coherence tomography slices

Patent metadata

Field	Value
Publication number	US-10219688-B2
Application number	US-201615296786-A
Country	US
Kind code	B2
Filing date	Oct 18, 2016
Priority date	Oct 19, 2015
Publication date	Mar 5, 2019
Grant date	Mar 5, 2019

Abstract

The present disclosure describes systems and methods to select fovea containing optical coherence tomography (OCT) images. The systems and methods described herein receive a plurality of OCT images. The portion of the OCT images are selected for further processing, where a line tracing the border between the retina and non-retina tissue is generated. A difference of the line is generated. Candidate OCT images are then generated responsive to the generated difference line. The lowest point among each difference lines generated for each of the OCT images is identified, and the OCT image to which the lowest point corresponds is identified as the fovea containing OCT image.

First claim

What is claimed:

1. A method for selecting optical coherence tomography (OCT) images, the method comprising:
receiving a plurality of optical coherence tomography (OCT) images;
for each of the plurality of OCT images:
identifying a boundary line between a first set of retina pixels and a second set of non-retina pixels in the respective OCT image;
calculating a derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image;

selecting an OCT image from the plurality of OCT images as containing a fovea by identifying the OCT image having the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels with a greatest magnitude; and
determining at least one feature of the retina tissue based on the selected OCT image.

2. The method of claim 1 , further comprising discarding a portion of the plurality of OCT images prior to identifying the boundary line between the first set of retina pixels and the second set of non-retina pixels.

3. The method of claim 1 , wherein the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels is a second derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels.

4. The method of claim 1 , further comprising determining if the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels crosses a predetermined threshold.

5. The method of claim 1 , further comprises receiving a prior probability distribution for the boundary line between the first set of retina pixels and the second set of non-retina pixels.

6. The method of claim 1 , further comprising calculating a probability that each point of the boundary line between the first set of retina pixels and the second set of non-retina pixels corresponds to a fovea.

7. The method of claim 6 , wherein the probability is calculated using Bayes' rule.

8. The method of claim 1 , further comprising identifying the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image using a Bayesian level set algorithm to classify each pixel of the respective OCT image as a retina or a non-retina pixel.

9. The method of claim 8 , wherein the boundary line between the first set of retina pixels and the second set of non-retina pixels is a substantially shortest path between the retina and the non-retina pixels.

10. The method of claim 1 , further comprising applying a smoothing filter to the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels.

11. A system for selecting clinically relevant optical coherence tomography (OCT) images, the system comprising a memory and one or more processors configured to execute instructions stored in the memory, execution of the instructions cause the one or more processors to:
receive a plurality of optical coherence tomography (OCT) images;
for each of the plurality of OCT images:
identify boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image;
calculate a derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image;

select an OCT image from the plurality of OCT images as containing a fovea by identifying the OCT image having the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image with a greatest magnitude; and
determine at least one feature of the retina tissue based on the selected OCT image.

12. The system of claim 11 , wherein the instructions cause the one or more processors to discard a portion of the plurality of OCT images prior to identifying the boundary line between the first set of retina pixels and the second set of non-retina pixels.

13. The system of claim 11 , wherein the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels is a second derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels.

14. The system of claim 11 , wherein the instructions cause the one or more processors to determine if the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels crosses a predetermined threshold.

15. The system of claim 11 , wherein the instructions cause the one or more processors to receive a prior probability distribution for the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image.

16. The system of claim 11 , wherein the instructions cause the one or more processors to calculate a probability that each point of the boundary line between the first set of retina pixels and the second set of non-retina pixels corresponds to a fovea.

17. The system of claim 16 , wherein the probability is calculated using Bayes' rule.

18. The system of claim 11 , wherein the instructions cause the one or more processors to identify the boundary line between the first set of retina pixels and the second set of non-retina pixels in the respective OCT image using a Bayesian level set algorithm to classify each pixel of the respective OCT image as a retina or a non-retina pixel.

19. The system of claim 18 , wherein the boundary line between the first set of retina pixels and the second set of non-retina pixels is a substantially shortest path between the retina and the non-retina pixels.

20. The system of claim 11 , wherein the instructions cause the one or more processors to apply a smoothing filter to the derivative of the boundary line between the first set of retina pixels and the second set of non-retina pixels.

Assignees

• Charles Stark Draper Laboratory Inc

Inventors

• Lowry Nathan
• Lommel Peter
• Hamilton Lei
• Mangoubi Rami

Classifications

• A61B3/1025 Primary

  for confocal scanning · CPC title

• A61B3/0025 Primary

  characterised by electronic signal processing, e.g. eye models · CPC title

• G06F18/24

  Classification techniques · CPC title

• A61B3/102

  for optical coherence tomography [OCT] · CPC title

• G06T2207/20076

  Probabilistic image processing · CPC title