Systems and methods to process electronic images to selectively hide structures and artifacts for digital pathology image review
A computer-implemented method for processing digital pathology images, the method including receiving a plurality of digital pathology images of at least one pathology specimen, the pathology specimen being associated with a patient. The method may further include determining, using a machine learning system, whether artifacts or objects of interest are present on the digital pathology images. Once the machine learning system has determined that an artifact or object of interest is present, the system may determine one or more regions on the digital pathology images that contain artifacts or objects of interest. Once the system determines the regions on the digital pathology images that contain artifacts or objects of interest, the system may use a machine learning system to inpaint or suppress the region and output the digital pathology images with the artifacts or objects of interest inpainted or suppressed.
1 . A computer-implemented method for processing digital pathology images, comprising:
receiving a plurality of digital pathology images of at least one pathology specimen, the pathology specimen being associated with a patient;
determining, using a first machine learning system, whether artifacts are present on the digital pathology images, the first machine learning system using artifact-agnostic learning techniques;
upon determining that an artifact is present, determining one or more regions on the digital pathology images that contain artifacts;
upon determining the one or more regions on the digital pathology images that contain artifacts, using a second machine learning system to inpaint or suppress the one or more regions; and
outputting the digital pathology images with the artifacts inpatined or suppressed.
2 . The method of claim 1 , wherein the artifact-agnostic learning techniques include applying classification-based learning techniques.
3 . The method of claim 2 , wherein applying the classification-based learning techniques to determine whether artifacts are presents further includes:
assigning a score to each patch within the plurality of digital pathology images; and
determining that the assigned scores are above a threshold value to classify a respective patch as including artifacts.
4 . The method of claim 1 , wherein the artifact-agnostic learning techniques include applying segmentation-based learning techniques.
5 . The computer-implemented method of claim 1 , wherein the artifacts includes writing utensil marking, hair, blur, scanlines, and/or bubbles displayed on the plurality of digital pathology images.
6 . The method of claim 1 , further comprising:
determining a segmentation map of the one or more regions on the digital pathology images that contain artifacts.
7 . A system for processing electronic medical images, the system comprising:
at least one memory storing instructions; and
at least one processor configured to execute the instructions to perform operations comprising:
receiving a plurality of digital pathology images of at least one pathology specimen, the pathology specimen being associated with a patient;
determining, using a first machine learning system, whether artifacts are present on the digital pathology images, the first machine learning system using artifact-agnostic learning techniques;
upon determining that an artifact is present, determining one or more regions on the digital pathology images that contain artifacts;
upon determining the one or more regions on the digital pathology images that contain artifacts, using a second machine learning system to inpaint or suppress the one or more regions; and
outputting the digital pathology images with the artifacts inpatined or suppressed.
8 . The system of claim 7 , wherein the artifact-agnostic learning techniques include applying classification-based learning techniques.
9 . The system of claim 8 , wherein applying the classification-based learning techniques to determine whether artifacts are presents further includes:
assigning a score to each patch within the plurality of digital pathology images; and
determining that the assigned scores are above a threshold value to classify a respective patch as including artifacts.
10 . The system of claim 7 , wherein the artifact-agnostic learning techniques include applying segmentation-based learning techniques.
11 . The system of claim 7 , wherein the artifacts includes writing utensil marking, hair, blur, scanlines, and/or bubbles displayed on the plurality of digital pathology images.
12 . The system of claim 7 , further comprising:
determining a segmentation map of the one or more regions on the digital pathology images that contain artifacts.
13 . A system for processing electronic medical images, the system comprising
at least one memory storing instructions; and
at least one processor configured to execute the instructions to perform operations comprising:
receiving a plurality of digital pathology images of at least one pathology specimen, the pathology specimen being associated with a patient;
determining, using a first machine learning system, whether artifacts are present on the digital pathology images, the first machine learning system using artifact-specific learning techniques;
receiving, from one or more users, a first artifact type to search for and remove;
upon determining that the first artifact type is present, determining one or more regions on the digital pathology images that contain the first artifact type;
upon determining the one or more regions on the digital pathology images that contain the first artifact type, using a second machine learning system to inpaint or suppress the one or more regions; and
outputting the digital pathology images with the first artifact type inpatined or suppressed.
14 . The system of claim 13 , wherein the artifacts includes writing utensil marking, hair, blur, scanlines, and/or bubbles displayed on the plurality of digital pathology images.
15 . The system of claim 13 , wherein the first machine learning system using the artifact-specific learning techniques applies learning techniques based on a shape of one or more artifacts.
16 . The system of claim 13 , wherein the first machine learning system using the artifact-specific learning techniques applies learning techniques based on an appearance of one or more artifacts.
17 . The system of claim 13 , wherein the first artifact type is blur.