Three dimensional corneal imaging with gabor-domain optical coherence microscopy
A system for non-contact imaging of corneal tissue stored in a viewing chamber using Gabor-domain optical coherence microscopy (GDOCM), wherein a 3D numerical flattening procedure is applied to the image data to produce an at least substantially artifact-free en face view of the endothelium.
1. A method of imaging a tissue sample, the method of comprising:
(a) imaging a portion of the tissue sample to create image data;
(b) applying a low pass filter to the image data to create a first filtered data;
(c) thresholding the first filtered data to create thresholded data;
(d) median filtering the thresholded data to reduce thresholding artifacts;
(e) automatically identifying target cell structures through iterative machine processing to reduce outliers to define a fitted surface; and
(f) adjusting an intensity projection of a slice adjacent the fitted surface to accommodate errors introduced by at least one of (i) the thresholding and (ii) median filtering to provide substantially artifact free en face view of the portion of the tissue sample.
2. The method of claim 1 , wherein the imaging includes imaging with Gabor-domain optical coherence microscopy (GDOCM).
3. The method of claim 1 , wherein applying a low pass filter to the image data includes applying a Gaussian blur to the image data.
4. The method of claim 1 , wherein the thresholding includes applying a binarization to the first filtered data.
5. The method of claim 4 , wherein the binarization is a Huang binarization.
6. The method of claim 1 , wherein the thresholding provides a bimodal acquired volume.
7. The method of claim 1 , wherein the iterative machine processing automatic identifying of target cells includes random sample consensus (RANSAC).
8. The method of claim 1 , further comprising applying a machine learning algorithm to the en face view.
9. The method of claim 1 , wherein the slice is below the fitted surface.
10. The method of claim 1 , wherein the portion of the tissue is curved.
11. The method of claim 1 , wherein the portion of the tissue is a scroll.
12. The method of claim 1 , wherein imaging the portion of the tissue sample to create the image data includes non-contact imaging the portion of the tissue sample to create image data.
13. The method of claim 1 , wherein the image data is three dimensional (3D) image data.
14. A method of imaging a tissue sample, the method of comprising:
(a) imaging a portion of the tissue sample with Gabor-domain optical coherence microscopy (GDOCM) to create image data;
(b) numerically flattening the image data; and
(c) displaying an artifact free en face image of the curved portion of the tissue.
15. The method of claim 14 , wherein numerically flattening includes applying a low pass filter to the image data to create a first filtered data, thresholding the first filtered data to create thresholded data, median filtering the thresholded data to reduce thresholding artifacts, automatically identifying target cell structures through iterative machine processing to reduce outliers to define a fitted surface and adjusting an intensity projection of a slice adjacent the fitted surface.
16. The method of claim 14 , further comprising counting cells in the en face image.
17. The method of claim 14 , wherein imaging a portion of the tissue sample includes non-contact imaging a curved area of the portion of the tissue sample.
18. An apparatus for imaging a tissue sample, the apparatus comprising:
(a) a Gabor-domain optical coherence microscope (GDOCM) imaging of a portion of the tissue sample; and
(b) a processor connected to the GDOCM, the processor configured to (i) apply a low pass filter to the image data to create a first filtered data; (ii) threshold the first filtered data to create thresholded data; (iii) median filtering the thresholded data to reduce thresholding artifacts; (iv) automatically identify target structures through iterative machine processing to reduce outliers to define a fitted surface; and (v) adjust an intensity projection of a slice adjacent the fitted surface to accommodate errors introduced by at least one of (aa) the image thresholding and (bb) median filtering to provide a substantially artifact free en face view of the portion of the tissue.
19. The method of claim 18 , wherein the Gabor-domain optical coherence microscope has a working distance that is greater than 0.0 mm for non-contact imaging the portion of the tissue sample.
20. A method of assessing tissue graft efficacy, the method comprising:
(a) in vivo imaging a grafted tissue with a Gabor-domain optical coherence microscope;
(b) applying a low pass filter to the image data to create a first filtered data;
(c) thresholding the first filtered data to create thresholded data;
(d) median filtering the thresholded data to reduce thresholding artifacts;
(e) automatically identifying target cells through iterative machine processing to reduce outliers to define a fitted surface; and
(f) adjusting an intensity projection of a slice adjacent the fitted surface to accommodate errors introduced by at least one of (aa) the image thresholding and (bb) median filtering to provide a substantially artifact free en face view of the grafter tissue.
21. The method of claim 20 , wherein the in vivo imaging includes non-contact imaging the grafted tissue.