Classifying image features
View Patent ↗Methods are disclosed for classifying different parts of a sample into respective classes based on an image stack that includes one or more images.
1. A method comprising using one or more processors to perform the following steps:
decomposing a set of spectral images of a sample into an unmixed image set, wherein each member of the unmixed image set corresponds to a spectral contribution from a different component in the sample; and
classifying different parts of the sample into respective classes based on an image stack comprising one or more of the unmixed images.
2. The method of claim 1 , wherein the images in the image stack used for classification comprise two or more of the unmixed images.
3. The method of claim 1 , wherein classification involves both spectral and spatial information in the image stack.
4. The method of claim 1 , further comprising:
generating a composite image based on the set of spectral images of the sample,
wherein the spatial intensities of two or more different members of a group consisting of the spectral images and the unmixed images are weighted differently and combined to produce the composite image, and
wherein the one or more images in the image stack comprise the composite image.
5. The method of claim 4 , wherein the set of spectral images comprises n images, and wherein the one or more images in the image stack used for classification comprises fewer than n images.
6. The method of claim 4 , wherein the composite image is generated according to a weighting function that changes monotonically with a spectral wavelength.
7. The method of claim 4 , wherein the spatial intensities of two or more different ones of the unmixed images are weighted differently and combined to produce the composite image.
8. The method of claim 4 , wherein the weighting function is selected to enhance a contrast between features contributed to the composite image from the two or more different members of the group consisting of the spectral images and the unmixed images.
9. The method of claim 1 , wherein a neural network is used for the classifying.
10. The method of claim 9 , wherein classifying different regions of the sample into the different classes comprises identifying selected regions of the image stack that correspond to each of the individual classes, training the neural network to recognize the classes based on the selected regions, and applying the trained neural network to the additional regions of the image stack.
11. The method of claim 1 , further comprising generating an output image showing classified regions of the sample.
12. The method of claim 1 , further comprising obtaining the one or more images in the image stack.
13. The method of claim 1 , wherein the sample is a tissue section.
14. Apparatus comprising a computer readable medium storing a program that causes a processor to carry out the method of claim 1 .
15. The method of claim 1 , further comprising outputting information about the classified different parts of the sample to a user.
16. The method of claim 1 , wherein the set of spectral images comprises n images, and wherein the one or more images in the image stack used for classification comprises fewer than n images.
17. The method of claim 1 , wherein the unmixed image set provides information about the relative amounts of the different components at each of multiple, different locations of the sample.
18. The method of claim 1 , wherein the components comprise absorbing stains or fluorescent labels.
19. The method of claim 1 , wherein the classes correspond to different biological or chemical structures in the sample.
20. The method of claim 1 , wherein the spectral images are optical spectral images.
21. Apparatus comprising:
a means for obtaining a set of spectral images of a sample; and
an electronic processor coupled to the means and configured to:
i) decompose the set of spectral images of a sample into an unmixed image set, wherein each member of the unmixed image set corresponds to a spectral contribution from a different component in the sample, and
ii) classify different parts of the sample into respective classes based on an image stack comprising one or more of the unmixed images.
22. The apparatus of claim 21 , further comprising an output device coupled to the electronic processor for outputting information about the classified different parts of the sample to a user.
23. The apparatus of claim 21 , wherein the set of spectral images comprises n images, and wherein the one or more images in the image stack used for classification comprises fewer than n images.
24. The apparatus of claim 21 , wherein the unmixed image set provides information about the relative amounts of the different components at each of multiple, different locations of the sample.
25. The apparatus of claim 21 , wherein the means for obtaining a set of spectral images is a means for obtaining a set of optical spectral images.
26. The apparatus of claim 21 , wherein the images in the image stack used for classification comprise two or more of the unmixed images.
27. The apparatus of claim 21 , wherein the classification by the electronic processor involves both spectral and spatial information in the image stack.
28. The apparatus of claim 21 , wherein the electronic processor uses a neural network for the classifying.
29. The apparatus of claim 28 , wherein the electronic processor is configured to classify different regions of the sample into the different classes by identifying selected regions of the image stack that correspond to each of the individual classes, training the neural network to recognize the classes based on the selected regions, and applying the trained neural network to the additional regions of the image stack.
30. The apparatus of claim method of claim 21 , wherein the electronic processor is configured to generate an output image showing classified regions of the sample.