# US 12,217,483 B2 · Systems and methods for processing electronic images for generalized disease detection > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,217,483 B2 - Document ID: US12217483B2 - Status: Granted - Kind: B2 - Application number: 18/488,364 - Filed: 2023-10-17 - Published: 2024-02-08 - Granted: 2025-02-04 - Earliest claimed: 2020-01-03 - Assignees: Paige.AI, Inc. - Applicants: PAIGE.AI, Inc. - Inventors: Belma Dogdas; Christopher Kanan; Thomas Fuchs; Leo Grady - Web page: https://patentgrep.com/patent/US12217483B2 ## CPC - G06V 10/764: using classification, e.g. of video objects - G06T 7/0012: {Biomedical image inspection} - G06V 10/82: using neural networks - G06V 20/698: {Matching; Classification} - G16H 30/40: for processing medical images, e.g. editing - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - G06T 2207/20081: Training; Learning - G06T 2207/30024: Cell structures in vitro; Tissue sections in vitro - G06T 2207/30096: Tumor; Lesion - G06V 2201/10: Recognition assisted with metadata - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA ## Abstract Systems and methods are disclosed for generating a specialized machine learning model by receiving a generalized machine learning model generated by processing a plurality of first training images to predict at least one cancer characteristic, receiving a plurality of second training images, the first training images and the second training images include images of tissue specimens and/or images algorithmically generated to replicate tissue specimens, receiving a plurality of target specialized attributes related to a respective second training image of the plurality of second training images, generating a specialized machine learning model by modifying the generalized machine learning model based on the plurality of second training images and the target specialized attributes, receiving a target image corresponding to a target specimen, applying the specialized machine learning model to the target image to determine at least one characteristic of the target image, and outputting the characteristic of the target image. ## Claims ### Claim 1 A computer-implemented method for processing electronic images, the method comprising: receiving a generalized machine learning model; receiving a plurality of training images, a quantity of training images being insufficient to generate a machine learning model that meets a threshold; receiving a plurality of target specialized attributes each related to a respective training image of the plurality of training images; generating a specialized machine learning model by modifying the generalized machine learning model based on the plurality of training images and respective target specialized attributes, the specialized machine learning model meeting the threshold, the specialized machine learning model being generated in accordance with a large-margin scheme built over one or more features of the generalized machine learning model; receiving a target image corresponding to a target specimen; applying the specialized machine learning model to the target image to determine at least one characteristic of the target image; and outputting the at least one characteristic of the target image. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, further comprising: determining a prediction of a specimen type of the target specimen based on the at least one characteristic of the target image; and outputting the prediction of the specimen type of the target specimen. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, wherein the plurality of target specialized attributes are one or more biomarkers present within each respective training image. ### Claim 4 (depends on 1) The computer-implemented method of claim 1, wherein the generalized machine learning model comprises a plurality of layers and modifying the generalized machine learning model further comprises modifying one or more outer layers of the generalized machine learning model. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, wherein modifying the generalized machine learning model further comprises removing an output layer of the generalized machine learning model. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein the plurality of target specialized attributes are one or more indications of characteristic outputs selected from a disease presence, staging variable presence, drug response, toxicity, or cancer classification. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, wherein the plurality of target specialized attributes are based on at least one of drug response information, cancer recurrence prediction information, or toxicity assessment information. ### Claim 8 (depends on 1) The computer-implemented method of claim 1, wherein the large-margin scheme reduces generalization. ### Claim 9 (depends on 1) The computer-implemented method of claim 1, wherein each of the training images are generated based on a same category of pathology specimens and wherein a category of pathology specimens is selected from histology, cytology, immunohistochemistry, or a combination thereof. ### Claim 10 (depends on 1) The computer-implemented method of claim 1, wherein modifying a generalized machine learning model further comprises adjusting the generalized machine learning model to have outputs based on the target specialized attributes. ### Claim 11 (depends on 1) The computer-implemented method of claim 1, wherein generalized machine learning model is generated by processing a plurality of first training images to predict at least one cancer characteristic. ### Claim 12 (depends on 1) The computer-implemented method of claim 1, wherein the at least one characteristic of the target image is one of a cancer diagnosis, a tumor characterization, or biomarker detection. ### Claim 13 A system comprising: at least one memory storing instructions; and at least one processor executing the instructions to perform operations comprising: receiving a generalized machine learning model; receiving a plurality of training images, a quantity of training images being insufficient to generate a machine learning model that meets a threshold; receiving a plurality of target specialized attributes each related to a respective training image of the plurality of training images; generating a specialized machine learning model by modifying the generalized machine learning model based on the plurality of training images and the respective target specialized attributes, the specialized machine learning model meeting the threshold, the specialized machine learning model being generated in accordance with large-margin scheme built over one or more features of the generalized machine learning model; receiving a target image corresponding to a target specimen; applying the specialized machine learning model to the target image to determine at least one characteristic of the target image; and outputting the at least one characteristic of the target image. ### Claim 14 (depends on 13) The system of claim 13, the operations further comprising: determining a prediction of a specimen type of the target specimen based on the at least one characteristic of the target image; and outputting the prediction of the specimen type of the target specimen. ### Claim 15 (depends on 13) The system of claim 13, wherein the large-margin scheme reduces generalization. ### Claim 16 (depends on 13) The system of claim 13, wherein the generalized machine learning model comprises a plurality of layers and modifying the generalized machine learning model further comprises modifying one or more outer layers of the generalized machine learning model. ### Claim 17 (depends on 13) The system of claim 13, wherein modifying the generalized machine learning model further comprises removing an output layer of the generalized machine learning model. ### Claim 18 A non-transitory computer-readable medium storing instructions that, when executed by processor, cause the processor to perform operations comprising: receiving a generalized machine learning model; receiving a plurality of training images, wherein a quantity of training images being insufficient to generate a machine learning model that meets a threshold; receiving a plurality of target specialized attributes each related to a respective training image of the plurality of training images; generating a specialized machine learning model by modifying the generalized machine learning model based on the plurality of training images and the respective target specialized attributes, the specialized machine learning model meeting the threshold, the specialized machine learning model being generated in accordance with a large margin scheme built over one or more features of the generalized machine learning model; receiving a target image corresponding to a target specimen; applying the specialized machine learning model to the target image to determine at least one characteristic of the target image; and outputting the at least one characteristic of the target image. ### Claim 19 (depends on 18) The non-transitory computer-readable medium of claim 18, the operations further comprising: determining a prediction of a specimen type of the target specimen based on the at least one characteristic of the target image; and outputting the prediction of the specimen type of the target specimen. ### Claim 20 (depends on 18) The non-transitory computer-readable medium of claim 18, wherein the plurality of target specialized attributes are one or more biomarkers present within each respective training image.