US 11,893,819 B2
Systems and methods for extracting and processing data using optical character recognition in real-time environments
Kenneth Cardozo, McLean, VA (US); Landon Nehmer, McLean, VA (US); Esmat Zare, McLean, VA (US); Mani Afsari, McLean, VA (US); Jitender Jain, McLean, VA (US); Venkateshwar Parpelli, McLean, VA (US); Bhuvaneswari Balasubramanian, McLean, VA (US); Bijun Du, McLean, VA (US); Daniel Nizinski, McLean, VA (US); and Tausif Shahid, McLean, VA (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Jul. 30, 2021, as Appl. No. 17/389,513.
Prior Publication US 2023/0034513 A1, Feb. 2, 2023
Int. Cl. G06V 30/416 (2022.01); G06N 3/08 (2023.01); G06N 5/01 (2023.01); G06V 30/19 (2022.01); G06V 10/75 (2022.01)
CPC G06V 30/416 (2022.01) [G06N 3/08 (2013.01); G06N 5/01 (2023.01); G06V 10/751 (2022.01); G06V 30/19007 (2022.01); G06V 30/19013 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system for automatically extracting and processing data using optical character recognition in real-time environments, the system comprising:
cloud-based storage circuitry configured to store an artificial intelligence model, wherein the artificial intelligence model is trained to perform optical character recognition to determine matches between text strings parsed from images and inputted target text strings;
cloud-based control circuitry configured to:
receive, from a user submission, a value to be verified, wherein the value corresponds to a field of a plurality of fields in a structured submission template, and wherein the user submission is based on a set of processing terms;
determining a target text string based on the value, wherein determining the target text string based on the value further comprises:
determining a first unit of measure corresponding to the field;
determining a second unit of measure corresponding to a respective field of a document, wherein the document comprises data for verifying the value;
determining a conversion metric for the target text string based on the first unit of measure and the second unit of measure; and
converting the value to the second unit of measure based on the conversion metric;
receiving an image of a document;
parsing the image to identify a plurality of text strings in the document;
comparing each of the plurality of text strings to the target text string;
determining that a text string of the plurality of text strings corresponds to the target text string based on comparing, using the artificial intelligence model, each of the plurality, of text strings to the target text string;
in response to determining that the text string of the plurality of text strings corresponds to the target text string, validating the user submission; and
in response to validating the user submission, processing the user submission to determine a recommendation based on the user submission, wherein processing the user submission further comprises:
submitting the user submission to a rules processor workflow, wherein the rules processor workflow comprises a decision tree for evaluating the user submission; and
receiving an output from the rules processor workflow, wherein the recommendation is based on the output;
cloud-based input/output circuitry configured to generate for display, in a user interface, the recommendation.