Systems and methods for processing electronic images to detect contamination

1. A computer-implemented method for processing an electronic image corresponding to a tissue specimen, the method comprising:

receiving, from at least one of one or more physician servers, one or more hospital servers, one or more clinical trial servers, and one or more research lab servers, a plurality of digital images associated with a tissue specimen;

detecting a plurality of image regions from a background of the digital images;

extracting a plurality of features from the plurality of image regions;

determining a prediction, using a machine learning system, of whether a first image region of the plurality of image regions is likely to comprise at least one external contaminant by comparing the plurality of features of the first image region of the image regions with the plurality of features of the other of the image regions, the machine learning system having been trained using a plurality of training images to predict a presence of external contaminants and/or a location of any external contaminants present in the tissue specimen; and

determining, based on the prediction of whether the first image region comprises the at least one external contaminant, whether to process the first image region using a processing algorithm.

2. The computer implemented method of claim 1 , wherein the prediction is determined based on at least one of a binary variable indicating an external contaminant is present, a numeric confidence score indicating presence of an external contaminant, a bounding box around a likely external contaminant and a confidence score for the bounding box, a segmentation mask for the external contaminant with a confidence score.

3. The computer implemented method of claim 1 , wherein the detecting of the plurality of image regions from the background of the plurality of digital images comprises:

thresholding based on color intensity, texture features, and/or Otsu's method.

4. The computer implemented method of claim 1 , wherein the detecting of the plurality of image regions from the background of the plurality of digital images comprises using one or more segmentation algorithms.

5. The computer implemented method of claim 1 , wherein the detecting of the plurality of image regions from the background of the plurality of digital images comprises using bounding box detection.

6. The computer implemented method of claim 1 , further comprising:

storing the plurality of digital images in at least one database along with a timestamp associated with the plurality of digital images.

7. The computer implemented method of claim 1 , wherein the machine learning system is implemented by a convolutional neural network (CNN), Region CNN (R-CNN), Faster R-CNN, Mask R-CNN, Convolutional Graph Neural Network, and/or Relationship Neural Network.

8. The computer implemented method of claim 1 , wherein the machine learning system is implemented by feature extraction using hand-engineered features and at least one of a random forest and deep learning, wherein the hand-engineered features comprise at least one from Scale-Invariant Feature Transform (SIFT), Speeded up robust features (SURF), Rotation Invariant Feature Transform (RIFT), Oriented feature from accelerated segment test (FAST) and/or Rotated Binary Robust Independent Elementary Features (BRIEF) (ORB).

9. The computer implemented method of claim 1 , wherein the machine learning system is trained using by plurality of training images by:

receiving the plurality of training images, wherein at least one of the plurality of training images includes at least one external contaminant and/or a background;

receiving information about a localization of each external contaminant in at least one training image of the plurality of training images;

receiving information about a tissue type of the tissue specimen depicted in at least one training image;

detecting one or more image regions from the background of the training image; and

training the machine learning system to indicate the presence of an external contaminant somewhere in the training image and/or a specific location of at least one external contaminant.

10. The computer implemented method of claim 1 , further comprising:

receiving information about a tissue type shown in the one or more digital images of the received one or more digital images; and

determining the prediction at least in part based on the tissue type.

11. A system for processing an electronic image corresponding to a tissue specimen, the system comprising:

at least one memory storing instructions; and

at least one processor configured to execute the instructions to perform operations comprising:

receiving from at least one of one or more physician servers, one or more hospital servers, one or more clinical trial servers, and one or more research lab servers, a plurality of digital images associated with a tissue specimen;

detecting a plurality of image regions from a background of the digital images;

extracting a plurality of features from the plurality of image regions;

determining a prediction, using a machine learning system, of whether a first image region of the plurality of image regions is likely to comprise at least one external contaminant by comparing the plurality of features of the first image region of the image regions with the plurality of features of the other of the image regions, the machine learning system having been trained using a plurality of training images to predict a presence of external contaminants and/or a location of any external contaminants present in the tissue specimen; and

determining, based on the prediction of whether the first image region comprises the at least one external contaminant, whether to process the image region using a processing algorithm.

12. The system of claim 11 , wherein a background of the one or more digital images comprises a portion of the image that is not the tissue specimen.

13. The system of claim 11 , wherein the detecting of the plurality of image regions from a background of the plurality of digital images comprises:

thresholding based on color intensity, texture features, and/or Otsu's method.

14. The system of claim 11 , wherein the detecting of the plurality of image regions from the background of the plurality of digital images comprises using one or more segmentation algorithms.

15. The system of claim 11 , wherein the detecting of the plurality of image regions from the background of the plurality of digital images comprises using bounding box detection.

16. The system of claim 11 , further comprising:

storing the plurality of digital images in at least one database along with a timestamp associated with the plurality of digital images.

17. The system of claim 11 , wherein the machine learning system is trained using by plurality of training images by:

receiving a plurality of training images, wherein at least one of the training images includes at least one external contaminant and/or a background;

receiving information about a localization of each external contaminant in at least one training image of the plurality of training images;

receiving information about a tissue type of the tissue specimen depicted in at least one training image;

detecting one or more image regions from the background of the training image; and

training a machine learning system to indicate the presence of an external contaminant somewhere in the training image and/or a specific location of at least one external contaminant.

18. The system of claim 11 , wherein the machine learning system is implemented by CNN, R-CNN, Faster R-CNN, Mask R-CNN, Convolutional Graph Neural Network, and/or Relationship Neural Network.

19. At least one non-transitory computer readable medium for processing an electronic image corresponding to a tissue specimen, the at least one non-transitory computer readable medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:

receiving, from at least one of one or more physician servers, one or more hospital servers, one or more clinical trial servers, and one or more research lab servers, a plurality of digital images associated with a tissue specimen;

detecting a plurality of image regions from a background of the digital images;

extracting a plurality of features from the plurality of image regions;

determining a prediction, using a machine learning system, of whether a first image region of the plurality of image regions is likely to comprise at least one external contaminant by comparing the plurality of features of the first image region of the image regions with the plurality of features of the other of the image regions, the machine learning system having been trained using a plurality of training images to predict a presence of external contaminants and/or a location of any external contaminants present in the tissue specimen; and

determining, based on the prediction of whether the first image region comprises the at least one external contaminant, whether to process the first image region using a processing algorithm.