Automated methods and systems for detecting cells in stained specimen images
A system and a method for unveiling poorly visible or lightly colored nuclei in an input image are disclosed. An input image is fed to a color deconvolution module for deconvolution into two color channels that are processed separately before being combined. The input image is deconvolved into two separate images: a stain image and a counter stain image. A complement of the stain image is generated in order to clearly reflect the locations of the poorly visible or light-colored nuclei. The complement image and the counter stain image are optionally normalized and then combined and segmented, to generate an output image with clearly defined nuclei. Alternatively, the complement of the stain image and the counter stain image can optionally be normalized, and then segmented prior to being combined to generate the output image.
1. A method comprising:
receiving an image depicting a specimen;
deconvolving the image into at least a first information channel and a second information channel;
generating a stain image along the first information channel;
generating a complement of the stain image;
generating a counter stain image along the second information channel;
combining the complement of the stain image and the counter stain image to generate a combined image;
processing the combined image using a segmentation operation to detect a plurality of image objects in the combined image, each of the plurality of image objects corresponding to a cell nucleus of the specimen; and
outputting the combined image with the detected plurality of image objects.
2. The method of claim 1 , wherein the first information channel includes a first color channel.
3. The method of claim 2 , wherein the second information channel includes a second color channel, wherein the second color channel is different from the first color channel.
4. The method of claim 1 , wherein the first information channel indicates an immunohistochemistry (IHC) stain.
5. The method of claim 1 , wherein the second information channel indicates a counterstain.
6. The method of claim 1 , further comprising:
normalizing the complement of the stain image to generate a normalized complement of the stain image; and
combining the normalized complement of the stain image and the counter stain image to generate the combined image.
7. The method of claim 6 , further comprising:
normalizing the counter stain image to generate a normalized counter stain image; and
combining the normalized complement of the stain image and the normalized counter stain image to generate the combined image.
8. An image analysis system comprising one or more processors and coupled to at least one memory storing instructions which, when executed by the one or more processors, cause the tissue analysis system to perform the operations of:
deconvolving a specimen image into at least a first information channel and a second information channel;
generating a stain image along the first information channel;
generating a counter stain image along the second information channel;
generating a complement of the stain image;
process the complement of the stain image using a segmentation operation to generate a first segmented nuclei image, wherein the first segmented nuclei image indicates a first set of image objects in the specimen image that correspond to a first set of nuclei;
process the counter stain image using the segmentation operation to generate a second segmented nuclei image, wherein the second segmented nuclei image indicates a second set of image objects in the specimen that correspond to a second set of nuclei; and
combining the first and second segmented nuclei images to generate a combined segmented nuclei image, wherein the combined segment nuclei image indicates bot the first and second set of image objects.
9. The image analysis system of claim 8 , wherein the instructions further cause the image analysis system to perform the operations of:
normalizing the complement of the stain image to generate a normalized complement of the stain image; and
processing the normalized complement of the stain image using the segmentation operation to generate a first normalized segmented nuclei image, wherein combining the first and second segmented nuclei images comprises combining the first normalized segmented nuclei image and the second segmented nuclei image.
10. The image analysis system of claim 9 , wherein the instructions further cause the image analysis system to perform the operations of:
normalizing the counter stain image to generate a normalized counter stain image; and
processing the normalized counter stain image using the segmentation operation to generate a second normalized segmented nuclei image, wherein combining the first normalized segmented nuclei image and the second segmented nuclei image comprises combining the first and second normalized segmented nuclei images.
11. The image analysis system of claim 8 , wherein the first information channel includes a first color channel.
12. The image analysis system of claim 11 , wherein the second information channel includes a second color channel, wherein the second color channel is different from the first color channel.
13. The image analysis system of claim 8 , wherein the first information channel indicates an immunohistochemistry (IHC) stain.
14. The image analysis system of claim 8 , wherein the second information channel indicates a counterstain.
15. A non-transitory computer-readable medium including instructions which, when executed by one or more processors of an image analysis system, cause the image analysis system to perform a plurality of operations for detecting nuclei in a specimen image, the operations comprising:
deconvolving the specimen image into at least a first information channel and a second information channel;
generating a stain image along the first information channel;
generating a complement of the stain image;
generating a counter stain image along the second information channel;
combining the complement of the stain image and the counter stain image to generate a combined image;
processing the combined image using a segmentation operation to detect a plurality of image objects in the combined image, each of the plurality of image objects corresponding to a cell nucleus of the specimen; and
outputting the combined image with the detected plurality of image objects.
16. The non-transitory computer-readable medium of claim 15 , wherein the operations further comprise:
normalizing the complement of the stain image to generate a normalized complement of the stain image; and
combining the normalized complement of the stain image and the counter stain image to generate the combined image.
17. The non-transitory computer-readable medium of claim 16 , wherein the operations further comprise:
normalizing the counter stain image to generate a normalized counter stain image; and
combining the normalized complement of the stain image and the normalized counter stain image to generate the combined image.
18. The non-transitory computer-readable medium of claim 15 , wherein the first information channel includes a first color channel.
19. The non-transitory computer-readable medium of claim 15 , wherein the first information channel indicates an immunohistochemistry (IHC) stain.
20. The non-transitory computer-readable medium of claim 15 , wherein the second information channel indicates a counterstain.