Method and apparatus for image-based eye tracking for retinal diagnostic or surgery device

1. An eye tracking method for determining a position of an eye or a part of an eye in an image of an image sequence by performing a comparison between said image and a reference image, said process including:

aligning a set of images of the eye or the part of the eye to a same retinal position by shifting one or more images of the set of images within one or more respective image planes;

computing an enhanced reference image based on a combination of said set of aligned images; and

determining said position of the eye or the part of the eye in said image of said image sequence by comparing said image of said image sequence and said enhanced reference image to yield a motion estimation between said reference image and said image of said sequence.

2. The method of claim 1 , further comprising:

combining the aligned set of images by averaging the aligned set of images to generate the enhanced reference image.

3. The method of claim 1 , further comprising:

using said retina position determined in said image sequence for tracking in connection with a diagnosis and/or surgical device.

4. The method of claim 3 , wherein said diagnosis and/or surgical device is one of the following:

an OCT apparatus;

a refractive surgery device.

5. The method of claim 1 , further comprising the following:

performing OCT spot trace rectification to eliminate a saturated OCT spot trace from an image in said sequence by replacing the saturated area by an interpolation in a corresponding area of unsaturated images and compensating illumination variations using one or more suitable filters.

6. The method of claim 1 , comprising:

performing a global motion estimation between said reference image and said image of said sequence.

7. The method of claim 6 , wherein

said global motion estimation is refined by-performing a motion estimation based on parts or landmarks of said reference image and said image of said image sequence.

8. The method of claim 1 , wherein said enhanced reference image is used for

OCT spot rectification.

9. The method of claim 1 further comprising:

aligning image sequences taken at different periods of time by aligning an enhanced reference corresponding to one of said sequences with a reference image or an enhanced reference image corresponding to said second sequence.

10. The method of claim 9 , wherein said reference image corresponding to said second sequence is a red-free image.

11. The eye tracking method of claim 1 in which aligning a set of images to the same retinal position comprises motion estimation.

12. The method of claim 1 , further comprising:

using said retina position determined in said image sequence for positional compensation in connection with a diagnosis and/or surgical device.

13. The method of claim 1 , further comprising the following:

performing OCT spot trace rectification to eliminate a saturated OCT spot trace from an image in said sequence by replacing the saturated area by an average intensity in a corresponding area of unsaturated images and compensating illumination variations using one or more suitable filters.

14. The method of claim 5 in which the one or more suitable filters comprise spatial spectral filters.

15. The method of claim 13 in which the one or more suitable filters comprise spatial spectral filters.

16. The method of claim 1 , comprising:

performing a motion estimation based on parts or landmarks of said reference image and said image of said image sequence.

17. The method of claim 1 , wherein said enhanced reference image is used for

deriving a spectrally matched filter for selecting one or more retina spectrum components having the highest SNR ratio.

18. An eye tracking system for determining a position of an eye or a part of an eye in an image of an image sequence by performing a comparison between said image and a reference image, said system including:

a computer comprising:

a module for aligning a set of images by re-positioning individual images in the set of images to a same retinal position;

a module for computing an enhanced reference image based on a combination of said set of aligned images;

a module for determining said position in said image of said image sequence by comparing said image of said image sequence and said enhanced reference image to yield a motion estimation between said enhanced reference image and said image of said sequence; and

a module for performing OCT spot trace rectification to eliminate a saturated OCT spot trace from an image in said sequence by replacing the saturated area by an interpolation in a corresponding area of unsaturated images or an average intensity in a corresponding area of unsaturated images.

19. The system of claim 18 , wherein the computer further comprises:

a module for aligning said set of images which are taken before said image sequence to the same retinal position;

a module for combining the aligned set of images by averaging the aligned set of images to generate the enhanced reference image.

20. The system of claim 18 , wherein the computer further comprises:

a module for using said retina position determined in said image sequence for tracking in connection with a diagnosis or surgical device.

21. The system of claim 20 , wherein said diagnosis or surgical device is one of the following:

an OCT apparatus;

a refractive surgery device.

22. The system of claim 18 , wherein the computer further comprises

a module for compensating illumination variations using one or more suitable filters.

23. The system of claim 18 , wherein the computer further comprises one or more of the following:

a module for performing a global motion estimation between said reference image and said image of said sequence;

a module for performing a motion estimation based on parts or landmarks of said reference image and said image of said image sequence.

24. The system of claim 23 , wherein

said global motion estimation is refined by said motion estimation based on parts or landmarks.

25. The system of claim 18 , wherein said enhanced reference image is used for the following:

OCT spot rectification or

deriving a spectrally matched filter for selecting one or more retina spectrum components having the highest SNR ratio.

26. The system of claim 18 , wherein the computer further comprises:

a module for aligning image sequences taken at different periods of time by aligning an enhanced reference corresponding to one of said sequences with a reference image or an enhanced reference image corresponding to said second sequence.

27. The system of claim 26 , wherein said reference image corresponding to said second sequence is a red-free image.

28. The system of claim 18 , further comprising:

a module for using said retina position determined in said image sequence for positional compensation in connection with a diagnosis or surgical device.

29. The system of claim 22 in which the one or more suitable filters comprise spatial spectral filters.

30. A computer program product recorded on a non-transitory computer-readable medium, the computer program product comprising computer instructions that, when executed causes a computer to:

align a set of images of an eye or a part of an eye to a same retinal position by shifting one or more images of the set of images within one or more respective image planes;

compute an enhanced reference image based on a combination of said set of aligned images;

after computing said enhanced reference image, obtain an image sequence, said image sequence including an image of the eye or the part of the eye; and

after obtaining said image sequence, determine a position of the eye or the part of the eye in said image of said image sequence by comparing said image of said image sequence and said enhanced reference image to yield a motion estimation between said reference image and said image of said image sequence.