Systems and methods for detection of structures and/or patterns in images
The subject disclosure presents systems and computer-implemented methods for automatic immune cell detection that is of assistance in clinical immune profile studies. The automatic immune cell detection method involves retrieving a plurality of image channels from a multi-channel image such as an RGB image or biologically meaningful unmixed image. A cell detector is trained to identify the immune cells by a convolutional neural network in one or multiple image channels. Further, the automatic immune cell detection algorithm involves utilizing a non-maximum suppression algorithm to obtain the immune cell coordinates from a probability map of immune cell presence possibility generated from the convolutional neural network classifier.
1. A computer-implemented method comprising:
extracting a patch extracted from a test image, the patch being generated around a candidate location of a detected structure in the test image, wherein the candidate location is determined by one or more of radial detection, ring detection, or foreground thresholding;
convolving and subsampling regions of the patch to generate a plurality of connections, until a fully connected layer is derived; and
generating at least one probability map of one or more cellular structures within the test image based on the fully connected layer,
wherein a color unmixing operation is applied to obtain a specific color channel of the test image, and the plurality of connections of the convolving and subsampling operations are configured based on potential biological information of the detected structure within the specific color channel of the test image.
2. The computer-implemented method of claim 1 , further comprising applying a local maximum finding method to the probability map of at least a portion of the test image to identify a particular pixel of the at least a portion of the test image that will be used as the location of the detected structure.
3. The computer-implemented method of claim 1 , further comprising training a convolutional neural network to obtain a probable location of the one or more cellular structures.
4. The computer-implemented method of claim 3 , further comprising separating the image into color channels corresponding to the one or more cellular structures in the image.
5. The computer-implemented method of claim 4 , wherein the color channels include at least a cellular structure channel and a background image structure channel.
6. A system comprising:
one or more data processors; and
a non-transitory computer readable storage medium containing instructions which, when executed on the one or more data processors, cause the one or more data processors to perform actions including:
extracting a patch extracted from a test image, the patch being generated around a candidate location of a detected structure in the test image, wherein the candidate location is determined by one or more of radial detection, ring detection, or foreground thresholding;
convolving and subsampling regions of the patch to generate a plurality of connections, until a fully connected layer is derived; and
generating at least one probability map of one or more cellular structures within the test image based on the fully connected layer,
wherein a color unmixing operation is applied to obtain a specific color channel of the test image, and the plurality of connections of the convolving and subsampling operations are configured based on a potential biological information of the detected structure within the specific color channel of the test image.
7. The system of claim 6 , wherein the actions further comprise applying a local maximum finding method to the probability map of at least a portion of the test image to identify a particular pixel of the at least a portion of the test image that will be used as the location of the detected structure.
8. The system of claim 6 , wherein the actions further comprise training a convolutional neural network to obtain a probable location of the one or more cellular structures.
9. The system of claim 8 , wherein the actions further comprise separating the image into color channels corresponding to the one or more cellular structures in the image.
10. The system of claim 9 , wherein the color channels include at least a cellular structure channel and a background image structure channel.
11. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, including instructions configured to cause one or more data processors to perform actions including:
extracting a patch extracted from a test image, the patch being generated around a candidate location of a detected structure in the test image, wherein the candidate location is determined by one or more of radial detection, ring detection, or foreground thresholding;
convolving and sub sampling regions of the patch to generate a plurality of connections, until a fully connected layer is derived; and
generating at least one probability map of one or more cellular structures within the test image based on the fully connected layer,
wherein a color unmixing operation is applied to obtain a specific color channel of the test image, and the plurality of connections of the convolving and subsampling operations are configured based on a potential biological information of the detected structure within the specific color channel of the test image.
12. The computer-program product of claim 11 , wherein the actions further comprise applying a local maximum finding method to the probability map of at least a portion of the test image to identify a particular pixel of the at least a portion of the test image that will be used as the location of the detected structure.
13. The computer-program product of claim 11 , wherein the actions further comprise training a convolutional neural network to obtain a probable location of the one or more cellular structures.
14. The computer-program product of claim 13 , wherein the actions further comprise separating the image into color channels corresponding to the one or more cellular structures in the image.
15. The computer-program product of claim 14 , wherein the color channels include at least a cellular structure channel and a background image structure channel.