Cross-validation of data derived from images captured during diagnostic sessions
AssigneeVerily Life Sciences LLC
InventorsRohan Ganeshananthan, Brian Levinstein
Life of this patent
Estimating the term…
Abstract
Introduced here are diagnostic platforms able to verify subjects included in images in a time- and resource-efficient manner. By assessing the degree of consistency between multiple characterizations of an attribute that are derived from different data, a diagnostic platform can determine the likelihood that an image includes a given subject. An attribute could be characterized based on metadata that accompanies an image captured during a diagnostic session, information provided by an individual as part of the diagnostic session, an output produced by a classification model upon being applied to the image, or any combination thereof.
Claims25
A computer-implemented method comprising:
- acquiring, by a processor, a Digital Imaging and Communications in Medicine (DICOM) data object that includes
- pixel data corresponding to an image generated during a diagnostic session, and
- context data specifying an attribute of the diagnostic session;
- parsing, by the processor, the context data to establish a first characterization of the attribute;
- selecting, by the processor based on the attribute, a classification model from a library of classification models;
- applying, by the processor, the classification model to the pixel data to produce a second characterization of the attribute;
- comparing, by the processor, the first and second characterizations of the attribute;
- computing, by the processor, a degree of similarity between the first and second characterizations of the attribute,
- wherein a higher degree of similarity corresponds to a higher likelihood that the first and second characterizations of the attribute are substantially identical;
- and
- based on a comparison of the degree of similarity to a threshold, generating, by the processor, a notification that indicates a potential discrepancy in the attribute, as separately determined based on the pixel data and the context data, exists.
A computer-implemented method comprising:
- acquiring, by a processor, a data object that includes
- pixel data corresponding to an image generated during a diagnostic session, and
- context data specifying an attribute of the diagnostic session;
- applying, by the processor, a classification model to the pixel data to produce a first characterization of the attribute;
- comparing, by the processor, the first characterization of the attribute to a second characterization of the attribute determined based on the context data, information provided by an individual as part of the diagnostic session, or any combination thereof;
- estimating, by the processor, a degree of consistency between the first and second characterizations of the attribute;
- and
- performing, by the processor, an action based on the degree of consistency.
An electronic device comprising:
- a memory that includes instructions for cross-checking information acquired from multiple sources to confirm the information is related to a single diagnostic session,
- wherein the instructions, when executed by a processor, cause the processor to:
- acquire a data object that includes
- pixel data corresponding to an image generated during a diagnostic session, and
- context data specifying an attribute of the diagnostic session;
- parse the context data to identify a first characterization of the attribute;
- apply a classification model to the pixel data to produce a second characterization of the attribute;
- acquire information provided by an individual as part of the diagnostic session;
- parse the information to identify a third characterization of the attribute;
- estimate a degree of consistency between the first, second, and third characterizations of the attribute;
- and
- perform an action based on the degree of consistency.
A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations comprising:
- acquiring a first image and accompanying first data that specifies at least one attribute of a first diagnostic session;
- applying a classification model to the first image to produce a first characterization of an attribute of a first individual captured in the first image;
- populating a first record to associate the first image and the first characterization with at least a portion of the accompanying first data;
- acquiring a second image and accompanying second data that specifies at least one attribute of a second diagnostic session;
- applying the classification model to the second image to produce a second characterization of the attribute of a second individual captured in the second image;
- comparing the first and second characterizations of the feature to determine whether the first and second individuals are the same person;
- and
- responsive to determining that the first and second individuals are the same person,
- populating a second record to associate the second image with the first image.
Dependencies are inferred from the claim text.