Method and apparatus for removing tissue autofluorescence
View Patent ↗Techniques for removing image autoflourescence from fluorescently stained biological images are provided herein. The techniques utilize non-negative matrix factorization that may constrain mixing coefficients to be non-negative. The probability of convergence to local minima is reduced by using smoothness constraints. The non-negative matrix factorization algorithm provides the advantage of removing both dark current and autofluorescence.
1. A method for reducing autofluorescence in an image of a biological material, comprising the steps of:
accessing image data from two or more images of the biological material, wherein at least one of the images is fluorescently stained for a marker of interest and wherein at least one of the images is a reference image;
analyzing the image data of the two or more images using a non-negative matrix factorization algorithm to generate an estimate of the autofluorescence in at least one of the two or more images; and
generating an output of a corrected image with reduced autofluorescence.
2. The method of claim 1 , comprising iteratively improving the estimate of the autofluorescence.
3. The method of claim 2 , wherein iteratively improving the estimate of the autofluorescence comprises smoothing with a Gaussian kernel.
4. The method of claim 1 , wherein the non-negative matrix factorization algorithm comprises an estimate for the contribution of dark current.
5. The method of claim 1 , wherein the non-negative matrix factorization algorithm uses as an input the camera exposure time for each of the two or more images.
6. The method of claim 1 , wherein analyzing the image data comprises employing an alternating constrained least squares approach.
7. The method of claim 1 , comprising acquiring the two or more images by a two-step image acquisition or an extra channel image acquisition.
8. A non-transitory computer-readable medium comprising instructions for:
accessing image data from two or more images of the biological material, wherein at least one of the images is fluorescently stained for a marker of interest and wherein at least one of the images is a reference image;
analyzing the image data of the two or more images using a non-negative matrix factorization algorithm to generate an estimate of the autofluorescence in at least one of the two or more images; and
generating an output of a corrected image with reduced autofluorescence.
9. The non-transitory computer-readable medium of claim 8 , comprising instructions for iteratively improving the estimate of the autofluorescence.
10. The non-transitory computer-readable medium of claim 9 , wherein the instructions for iteratively improving the estimate of the autofluorescence comprise smoothing with a Gaussian kernel.
11. The non-transitory computer-readable medium of claim 8 , wherein the non-negative matrix factorization algorithm comprises an estimate for the contribution of dark current.
12. The non-transitory computer-readable medium of claim 8 , wherein the non-negative matrix factorization algorithm uses as an input the camera exposure time for each of the two or more images.
13. The non-transitory computer-readable medium of claim 8 , wherein the instructions for analyzing the image data comprise employing an alternating constrained least squares approach.
14. The non-transitory computer-readable medium of claim 8 , comprising instructions for acquiring the two or more images by a two-step image acquisition, or an extra channel image acquisition.
15. An image analysis system comprising:
a processor adapted to receive image data from two or more images of the biological material, wherein at least one of the images is fluorescently stained for a marker of interest and wherein at least one of the images is a reference image the processor adapted to run instructions for:
analyzing the image data of the two or more images using a non-negative matrix factorization algorithm to generate an estimate of the autofluorescence in at least one of the two or more images; and
generating an output of a corrected image with reduced autofluorescence.
16. The system of claim 15 , comprising instructions for iteratively improving the estimate of the autofluorescence.
17. The system of claim 16 , wherein the instructions for iteratively improving the estimate of the autofluorescence comprise smoothing with a Gaussian kernel.
18. The system of claim 15 , wherein the non-negative matrix factorization algorithm comprises an estimate for the contribution of dark current.
19. The system of claim 15 , wherein the non-negative matrix factorization algorithm uses as an input the camera exposure time for each of the two or more images.
20. The system of claim 15 , wherein the instructions for analyzing the image data comprise employing an alternating constrained least squares approach.
21. The system of claim 15 , comprising instructions for acquiring the two or more images by a two-step image acquisition or an extra channel image acquisition.