Systems and methods for processing electronic images to detect contamination in specimen preparations
Systems and methods are disclosed for receiving one or more digital images associated with a tissue specimen, detecting one or more image regions from a background of the one or more digital images, determining a prediction, using a machine learning system, of whether at least one first image region of the one or more image regions comprises at least one external contaminant, 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 a first image region comprises an external contaminant, whether to process the image region using an processing algorithm.
1. A computer-implemented method for processing digital images of a tissue specimen, the method comprising:
receiving, by a floater detection platform, the digital images of the tissue specimen;
receiving, by the floater detection platform, data associated with a tissue type of the tissue specimen, the data including a tissue class;
detecting, by the floater detection platform, one or more contiguous tissue pieces of the tissue specimen using an analysis of the digital images;
extracting, by the floater detection platform, a plurality of features from each contiguous tissue piece of the one or more contiguous tissue pieces of the tissue specimen;
providing, by the floater detection platform, the plurality of features from each contiguous tissue piece to a machine-learning model, wherein the machine-learning model has been trained, using one or more gathered and/or simulated sets of digital images of a plurality of tissue specimens, to identify a sample tissue class of each contiguous tissue piece based on the plurality of features;
outputting, by the machine-learning model, a first sample tissue class of a first contiguous tissue piece;
comparing, by the floater detection platform, the first sample tissue class of the first contiguous tissue piece with the tissue class of the tissue specimen to determine a tissue class mismatch; and
generating, by the floater detection platform, an alert based on the tissue class mismatch.
2. The computer-implemented method of claim 1 , wherein the analysis of the digital images is implemented by one or more of Faster R-CNN, selective search, EdgeBoxes, DeepBox, Mask R-CNN, or U-Net.
3. The computer-implemented method of claim 1 , wherein the plurality of features from each contiguous tissue piece comprise one or more of convolutional neural network (CNN) features, SIFT features, or SURF features.
4. The computer-implemented method of claim 1 , wherein the alert comprises one or more of an output transmitted to a computing device or a marker indicating a location of an external contaminant.
5. The computer-implemented method of claim 1 , wherein the analysis of the digital images comprises:
detecting, by the floater detection platform, a first image region from a background of a digital image of the digital images.
6. The computer-implemented method of claim 1 , further comprising:
providing, by the floater detection platform, the digital images to a database for storage, each digital image associated with a timestamp.
7. The computer-implemented method of claim 1 , wherein the machine-learning model 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 model is implemented by feature extraction using hand-engineered features and at least one of random forest, support vector machine (SVM), and/or deep learning.
9. The computer-implemented method of claim 8 , wherein the hand-engineered features comprise at least one of 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).
10. The computer-implemented method of claim 1 , wherein the floater detection platform includes a slide analysis tool.
11. A computer-implemented method for training a machine-learning model for processing digital images of a tissue specimen, the method comprising:
providing, by one or more processors, one or more gathered and/or simulated sets of digital images of a plurality of tissue specimens to one or more machine-learning algorithms as one or more sets of training data;
determining, by the one or more machine-learning algorithms, associations between the one or more gathered and/or simulated sets of digital images and one or more tissue features;
modifying one or more of a layer, a weight, a synapse, or a node of the machine-learning model based on the associations between the one or more gathered and/or simulated sets of digital images and the one or more tissue features; and
outputting, by the one or more processors, the machine-learning model, wherein the machine-learning model is trained to identify a sample tissue class of a contiguous tissue piece based on the one or more tissue features and the modified one or more of the layer, the weight, the synapse, or the node of the machine-learning model.
12. The computer-implemented method of claim 11 , wherein the machine-learning model 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.
13. The computer-implemented method of claim 11 , wherein the machine-learning model is implemented by feature extraction using hand-engineered features and at least one of random forest, support vector machine (SVM), and/or deep learning.
14. The computer-implemented method of claim 11 , further comprising:
providing, by the one or more processors, data associated with a tissue type of a tissue specimen, the data including a tissue class.
15. A system for processing digital images of a tissue specimen, the system comprising:
a memory storing instruction and a trained machine-learning model trained to identify a presence of one or more external contaminants and/or a location of the one or more external contaminants present and output an external contaminant prediction; and
a processor operatively connected to the memory and configured to execute the instructions to perform operations including:
receiving, by a floater detection platform, the digital images of the tissue specimen;
receiving, by the floater detection platform, data associated with a tissue type of the tissue specimen, the data including a tissue class;
detecting, by the floater detection platform, one or more contiguous tissue pieces of the tissue specimen using an analysis of the digital images;
extracting, by the floater detection platform, a plurality of features from each contiguous tissue piece of the one or more contiguous tissue pieces of the tissue specimen;
providing, by the floater detection platform, the plurality of features from each contiguous tissue piece to a machine-learning model, wherein the machine-learning model has been trained, using one or more gathered and/or simulated sets of digital images of a plurality of tissue specimens, to identify a sample tissue class of each contiguous tissue piece based on the plurality of features;
outputting, by the machine-learning model, a first sample tissue class of a first contiguous tissue piece;
comparing, by the floater detection platform, the first sample tissue class of the first contiguous tissue piece with the tissue class of the tissue specimen to determine a tissue class mismatch; and
generating, by the floater detection platform, an alert based on the tissue class mismatch.
16. The system of claim 15 , wherein the analysis of the digital images is implemented by one or more of Faster R-CNN, selective search, EdgeBoxes, DeepBox, Mask R-CNN, or U-Net.
17. The system of claim 15 , wherein the plurality of features from each contiguous tissue piece comprise one or more of convolutional neural network (CNN) features, SIFT features, or SURF features.
18. The system of claim 15 , wherein the alert comprises one or more of an output transmitted to a computing device or a marker indicating a location of an external contaminant.
19. The system of claim 15 , wherein the analysis of the digital images comprises:
detecting, by the floater detection platform, a first image region from a background of a digital image of the digital images.
20. The system of claim 15 , further comprising:
providing, by the floater detection platform, the digital images to a database for storage, each digital image associated with a timestamp.