# US 11,907,306 B2 · Systems and methods for classifying documents > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,907,306 B2 - Document ID: US11907306B2 - Status: Granted - Kind: B2 - Application number: 17/568,004 - Filed: 2022-01-04 - Published: 2023-07-06 - Granted: 2024-02-20 - Earliest claimed: 2022-01-04 - Assignees: Capital One Services, LLC - Applicants: Capital One Services, LLC - Inventors: Aaron Attar - Web page: https://patentgrep.com/patent/US11907306B2 ## CPC - G06F 16/906: Clustering; Classification - G06F 16/93: Document management systems - G06V 10/82: using neural networks - G06V 30/14: Image acquisition - G06V 30/1444: {Selective acquisition, locating or processing of specific regions, e.g. highlighted text, fiducial marks or predetermined fields} - G06V 30/18: Extraction of features or characteristics of the image - G06V 30/19: Recognition using electronic means - G06V 30/413: Classification of content, e.g. text, photographs or tables - G06V 30/418: Document matching, e.g. of document images - G06V 30/42: based on the type of document - G06V 30/416: Extracting the logical structure, e.g. chapters, sections or page numbers; Identifying elements of the document, e.g. authors - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A system may iteratively scan a portion of a document, extract first data from the portion of the document, and determine, using a trained model, whether the first data corresponds to one or more document types based on one or more confidence thresholds. The system may repeat this process, increasing the portion of the document scanned by a predetermined amount each iteration, until the first data corresponds to the one or more document types based on the one or more confidence thresholds. Responsive to determining the first data corresponds to the one or more document types based on the one or more confidence thresholds, the system may cause a graphical user interface (GUI) of a user device to display a notification indicating a document type match. ## Claims ### Claim 1 A system comprising: one or more processors; and a memory in communication with the one or more processors and storing instructions that, when executed by the one or more processors, are configured to cause the system to: receive a first single-page document; iteratively, until a document type is identified: scan a portion of the first single-page document, wherein the portion is a fraction of the first single-page document; extract first text data from the portion of the first single-page document by performing optical character recognition (OCR) on the portion of the first single-page document; and determine whether the document type is identified by, using a neural network (NN), determining whether the first text data corresponds to one or more document types of a plurality of document types based on one or more confidence thresholds, wherein the portion of the first single-page document scanned increases by a predetermined amount each iteration, the predetermined amount being dynamically updated each iteration by: determining whether a predefined amount of empty space is identified within the portion of the first single-page document; responsive to determining the predefined amount of empty space is identified within the portion of the first single-page document, increasing the portion of the first single-page document scanned by a second predetermined amount each iteration until second text data is identified; and responsive to the second text data being identified, automatically reverting to scanning the first single-page document based on an initial predetermined amount with each iteration; and responsive to determining the first text data corresponds to the one or more document types based on the one or more confidence thresholds, cause a graphical user interface (GUI) of a user device to display a first notification indicating a first document type match. ### Claim 2 (depends on 1) The system of claim 1, wherein the predetermined amount comprises a percentage. ### Claim 3 (depends on 1) The system of claim 1, wherein the instructions are further configured to cause the system to: stop iteratively scanning the portion of the first single-page document when the predetermined amount reaches a maximum limitation. ### Claim 4 (depends on 1) The system of claim 1, wherein each of the one or more confidence thresholds indicates a respective likelihood that the first text data corresponds to the one or more document types. ### Claim 5 A system comprising: one or more processors; and a memory in communication with the one or more processors and storing instructions that, when executed by the one or more processors, are configured to cause the system to: receive a first single-page document; iteratively, until a document type is identified: scan a portion of the first single-page document, wherein the portion is a fraction of the first single-page document; extract first data from the portion of the first single-page document; and determine whether the document type is identified by, using a machine learning model (MLM), determining whether the first data corresponds to one or more document types of a plurality of document types based on one or more confidence thresholds, wherein the portion of the first single-page document scanned increases by a predetermined amount each iteration, the predetermined amount being dynamically updated each iteration by: determining that a predefined amount of empty space is identified within the portion of the first single-page document; responsive to determining the predefined amount of empty space is identified within the portion of the first single-page document, increasing the portion of the first single-page document scanned by a second predetermined amount each iteration until second text data is identified; and responsive to the second text data being identified, automatically reverting to scanning the first single-page document based on an initial predetermined amount with each iteration; and responsive to determining the first data corresponds to the one or more document types based on the one or more confidence thresholds, transmit a first notification indicating a first document type match. ### Claim 6 (depends on 5) The system of claim 5, wherein the first data comprises text data, and wherein extracting the first data from the portion of the first single-page document is conducted by performing optical character recognition (OCR) on the portion of the first single-page document. ### Claim 7 (depends on 5) The system of claim 5, wherein the MLM comprises a neural network (NN). ### Claim 8 (depends on 5) The system of claim 5, wherein the predetermined amount comprises a percentage. ### Claim 9 (depends on 5) The system of claim 5, wherein the instructions are further configured to cause the system to: stop iteratively scanning the portion of the first single-page document when the predetermined amount reaches a maximum limitation. ### Claim 10 (depends on 5) The system of claim 5, wherein the instructions are further configured to cause the system to: cause a graphical user interface (GUI) of a user device to display the first notification. ### Claim 11 A system comprising: one or more processors; and a memory in communication with the one or more processors and storing instructions that, when executed by the one or more processors, are configured to cause the system to: receive a first single-page document; scan a first portion of the first single-page document, the first portion based on a first threshold and comprising a first fraction of the first single-page document; extract first data from the first portion of the first single-page document; determine, using a classifier, whether the first data corresponds to one or more document types of a plurality of document types based on one or more confidence thresholds; responsive to determining the first data corresponds to the one or more document types based on the one or more confidence thresholds, cause a graphical user interface (GUI) of a user device to display a first notification indicating a first document type match; and responsive to determining the first data does not correspond to the one or more document types based on the one or more confidence thresholds: iteratively, until a second document type is identified: scan a second portion of the first single-page document, wherein the second portion is a second fraction of the first single-page document; extract second data from the second portion of the first single-page document, the second portion comprising the first portion and based on a second threshold; determine, using the classifier, whether the second data corresponds to the one or more document types of the plurality of document types based on the one or more confidence thresholds, wherein the second portion increases by a dynamically updated predetermined amount each iteration; and responsive to determining the second data corresponds to the one or more document types based on the one or more confidence thresholds, cause the GUI of the user device to display a second notification indicating a second document type match. ### Claim 12 (depends on 11) The system of claim 11, wherein the first threshold comprises a first percentage and the second threshold comprises a second percentage. ### Claim 13 (depends on 12) The system of claim 12, wherein the first percentage is less than the second percentage. ### Claim 14 (depends on 12) The system of claim 12, wherein the first and second thresholds further comprise a top section or a bottom section of the first single-page document. ### Claim 15 (depends on 11) The system of claim 11, wherein the instructions are further configured to cause the system to: responsive to determining the first data does not correspond to the one or more document types based on the one or more confidence thresholds: cause the GUI of the user device to display a third notification indicating an error. ### Claim 16 (depends on 11) The system of claim 11, wherein the instructions are further configured to cause the system to: responsive to determining the second data does not correspond to the one or more document types based on the one or more confidence thresholds: scan a third portion of the first single-page document; extract third data from the third portion of the first single-page document, the third portion comprising the first portion and the second portion and based on a third threshold; determine, using the classifier, whether the third data corresponds to the one or more document types of the plurality of document types based on the one or more confidence thresholds; and responsive to determining the third data corresponds to the one or more document types based on the one or more confidence thresholds, cause the GUI of the user device to display a third notification indicating a third document type match. ### Claim 17 (depends on 16) The system of claim 16, wherein the first, the second, and the third thresholds are less than or equal to a maximum limitation. ### Claim 18 (depends on 11) The system of claim 11, wherein the classifier comprises a machine learning model (MLM). ### Claim 19 (depends on 18) The system of claim 18, wherein the MLM comprises a neural network (NN). ### Claim 20 (depends on 11) The system of claim 11, wherein the first data and the second data comprise text data.