Removal of non-retinal opthalmic reflections in retinal imaging

Patent metadata

Field	Value
Publication number	US-11013406-B2
Application number	US-201815962183-A
Country	US
Kind code	B2
Filing date	Apr 25, 2018
Priority date	Sep 4, 2009
Publication date	May 25, 2021
Grant date	May 25, 2021

Abstract

System and Method pertaining to the modification and integration of an existing consumer digital camera, for example, with an optical imaging module to enable point and shoot fundus photography of the eye. The auto-focus macro capability of existing consumer cameras is adapted to photograph the retina over an extended diopter range, eliminating the need for manual diopter focus adjustment. The thru-the-lens (TTL) auto-exposure flash capability of existing consumer cameras is adapted to photograph the retina with automatic flash exposure eliminating the need for manual flash adjustment. The consumer camera imaging sensor and flash are modified to allow the camera sensor to perform both non-mydriatic focusing of the retina using infrared illumination and standard color flash photography of the retina without the need for additional imaging sensors or mechanical filters. These modifications and integration of existing consumer cameras for fundus photography of the eye significantly improve ease of manufacture and usability.

First claim

We claim:

1. A method for removing non-retinal ophthalmic reflections from a retinal image, the method comprising:
obtaining information comprising a first retinal image from a camera device;
identifying a non-retinal ophthalmic reflection in the information comprising the first retinal image by automatically determining differences in the optical properties of the non-retinal ophthalmic reflection from the first retinal image compared to the optical properties of the first retinal image that do not contain the non-retinal ophthalmic reflection;
using the identified non-retinal ophthalmic reflection, automatically constructing a digital mask that corresponds to areas of the first retinal image containing the non-retinal ophthalmic reflection; and
removing the non-retinal ophthalmic reflection by digitally applying the mask to the first retinal image to provide a digitally-masked retinal image.

2. The method of claim 1 , wherein removing the non-retinal ophthalmic reflection by digitally applying the mask to the first retinal image includes subtracting values in the digital mask from values in the first retinal image at corresponding locations in the digital mask and the first retinal image.

3. The method of claim 1 , wherein removing the non-retinal ophthalmic reflection by digitally applying the mask to the first retinal image includes subtracting values corresponding to a spectral signature of the non-retinal ophthalmic reflection from the first retinal image.

4. The method of claim 1 , wherein identifying the non-retinal ophthalmic reflection in the first retinal image includes using information about at least one optical property comprising a spectral wavelength content, a size, a brightness, a contrast, or a hue of a region containing the non-retinal reflection as compared to at least one corresponding optical property of the first retinal image that lacks the non-retinal ophthalmic reflection.

5. The method of claim 4 , wherein identifying the non-retinal ophthalmic reflection in the first retinal image includes using pixel information in one or more color channels of the first retinal image.

6. The method of claim 5 , wherein the pixel information is used from a blue channel of the first retinal image.

7. The method of claim 1 , wherein the non-retinal ophthalmic reflection comprises Purkinje reflections from a cornea and a lens of an eye being imaged by the camera device.

8. The method of claim 1 , wherein the non-retinal ophthalmic reflection comprises scattered light from a cornea and a lens of an eye being imaged by the camera device.

9. The method of claim 1 , wherein the non-retinal ophthalmic reflection comprises light scattered within the camera device.

10. The method of claim 1 , wherein the non-retinal ophthalmic reflection comprises reflections from imaging optics within the camera device.

11. The method of claim 1 , wherein the identifying the non-retinal ophthalmic reflection, automatically constructing the digital mask, and removing the non-retinal ophthalmic reflection by digitally applying the mask to the first retinal image are performed remotely with respect to the camera device.

12. The method of claim 1 , wherein the identifying the non-retinal ophthalmic reflection, automatically constructing the digital mask, and removing the non-retinal ophthalmic reflection by digitally applying the mask to the first retinal image are performed using a processor included as a portion of the camera device.

13. A method to remove non-retinal ophthalmic reflection from at least one retinal image amongst a plurality of retinal images and to create a blended composite of the retinal images, the method comprising:
obtaining information comprising a plurality retinal images of a subject from a camera device, the plurality of retinal images obtained under similar illumination and corresponding to regions that overlap;
identifying a non-retinal ophthalmic reflection in at least one image amongst the plurality of retinal images by automatically determining differences in the optical properties of the non-retinal ophthalmic reflection from the at least one image compared to a retinal ophthalmic reflection;
using the identified non-retinal ophthalmic reflection, identifying a second image amongst the plurality of images that does not contain the non-retinal ophthalmic reflection in the same region;
automatically constructing a digital mask that corresponds to the region of the at least one retinal image containing the non-retinal ophthalmic reflection;
removing the non-retinal ophthalmic reflection by digitally applying data from the second image in the same region defined by the mask to the at least one retinal image to provide a digitally-masked retinal image; and
aggregating the retinal images including the digitally-masked retinal image to generate a composite for presentation to a user.

14. The method of claim 13 , wherein removing the non-retinal ophthalmic reflection by digitally applying the mask to at least one the retinal image includes subtracting values in the digital mask from values in the at least one retinal image at corresponding locations in the digital mask and the at least one retinal image.

15. The method of claim 13 , wherein removing the non-retinal ophthalmic reflection by digitally applying the mask to the at least one retinal image includes subtracting values corresponding to a spectral signature of the non-retinal ophthalmic reflection from the at least one retinal image.

16. The method of claim 13 , wherein identifying the non-retinal ophthalmic reflection in the at least one retinal image includes using information about at least one optical property comprising a spectral wavelength content, a size, a brightness, a contrast, or a hue of a region containing the non-retinal reflection as compared to at least one corresponding optical property of the at least one retinal image that lacks the non-retinal ophthalmic reflection.

17. The method of claim 13 , wherein identifying the non-retinal ophthalmic reflection in the at least one retinal image includes using pixel information in one or more color channels of the at least one retinal image.

18. The method of claim 17 , wherein the pixel information is used from a blue channel of the at least one retinal image.

19. The method of claim 13 , wherein an area of the at least one retinal image containing the non-retinal ophthalmic reflection is replaced in the composite with a corresponding overlapping area from at least one other retinal image that lacks the non-retinal ophthalmic reflection amongst the plurality of retinal images.

20. The method of claim 13 , comprising adjusting an exposure value used for acquisition of the at least one retinal image to provide even illumination across the composite when the retinal image is aggregated.

Assignees

• Univ Virginia Patent Foundation

Inventors

• Yates Paul Andrew
• Tran Kenneth

Classifications

• G06V40/19

  Sensors therefor · CPC title

• A61B3/156 Primary

  for blocking · CPC title

• A61B3/12

  for looking at the eye fundus, e.g. ophthalmoscopes (A61B3/13 takes precedence) · CPC title

• A61B3/14

  Arrangements specially adapted for eye photography · CPC title

• A61B3/1208

  Multiple lens hand-held instruments · CPC title