# US 12,216,690 B1 · Automated indexing and extraction of fields in digital records > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,216,690 B1 - Document ID: US12216690B1 - Status: Granted - Kind: B1 - Application number: 18/543,313 - Filed: 2023-12-18 - Granted: 2025-02-04 - Earliest claimed: 2023-01-17 - Assignees: VelocityEHS Holdings, Inc. - Applicants: VelocityEHS Holdings, Inc. - Inventors: Julia Penfield; Aatish Suman; Veeru Talreja; Misbah Zahid Khan - Web page: https://patentgrep.com/patent/US12216690B1 ## CPC - G06F 16/328: {Management therefor} - G06F 40/205: Parsing - G06F 40/258: Heading extraction; Automatic titling; Numbering - G06F 40/279: Recognition of textual entities - G06F 40/284: Lexical analysis, e.g. tokenisation or collocates - G06F 40/295: Named entity recognition - G06F 40/30: Semantic analysis - G06V 30/19: Recognition using electronic means - G06V 30/412: Layout analysis of documents structured with printed lines or input boxes, e.g. business forms or tables - G06V 30/416: Extracting the logical structure, e.g. chapters, sections or page numbers; Identifying elements of the document, e.g. authors - G06V 10/82: using neural networks - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract Systems and Methods are disclosed herein for automatically indexing multiple informational fields in digital data records, the method comprising: identifying, based on rules defining target information fields, for each target field of the target information fields, at least one page in a digital data record comprising content related to the target field; extracting, for each target field, from the identified at least one page, at least one portion of text comprising the content; feeding, for each target field, a pre-processed version of the at least one portion of text into a machine learning (ML) model, wherein the ML model is trained on the target field; determining, for each target field, via the ML model trained on the target field, at least one candidate text comprising the content; and extracting, for each target field, the at least one candidate text. ## Claims ### Claim 1 A method to automatically index multiple informational fields in digital data records, the method comprising: identifying, based on rules defining target fields, for each target field of the target fields, at least one page in a digital data record, the page comprising content related to the target field; extracting, for each target field, from the at least one page, at least one portion of text; inputting the at least one portion of text into a machine learning (ML) model, wherein the ML model is trained on the target field; determining, for the target field, via the ML model, at least one candidate text comprising the content; and extracting, the at least one candidate text. ### Claim 2 (depends on 1) The method of claim 1, wherein the at least one candidate text is empty and does not contain alphanumerical characters, symbols or other characters. ### Claim 3 (depends on 1) The method of claim 1, further comprising: identifying, based on rules defining the content, the at least one portion of text comprising the content within the at least one page. ### Claim 4 (depends on 1) The method of claim 1, further comprising: validating the at least one candidate text. ### Claim 5 (depends on 1) The method of claim 1, further comprising: generating, for each target field, an index based on the at least one candidate text. ### Claim 6 (depends on 4) The method of claim 4, further comprising: displaying the at least one candidate text for each target field on a graphical user interface. ### Claim 7 (depends on 1) The method of claim 1, further comprising: combining, the at least one candidate text from each target field. ### Claim 8 (depends on 1) The method of claim 1, further comprising: saving, for each target field, the at least one candidate text into a database. ### Claim 9 (depends on 1) The method of claim 1, wherein the extracting of the at least one portion of text comprising the content comprises: converting for each target field, from the at least one page, the at least one portion of text comprising the content from an image format into a text format. ### Claim 10 (depends on 9) The method of claim 9, wherein the text format is a string. ### Claim 11 (depends on 1) The method of claim 1, further comprising: identifying a page of the digital data record, the page containing special target information; inputting an image of the page into a visual machine learning network (visual ML), wherein the visual ML is trained to recognize text associated with the special target information in an image; identifying by the visual ML, a section of the image that contains the special target information; inputting at least one of a page number of the page, the digital data record, or coordinates of the section into an extraction module; and extracting the special target information by the extraction module from the section. ### Claim 12 (depends on 11) The method of claim 11, further comprising: inputting the special target information into a natural language processing ML network (NLP); and identifying at least one data item, by the NLP, based on a presence of at least one data item in the special target information, and a structure of the special target information. ### Claim 13 A system to automatically index multiple informational fields in a digital data record, the system comprising: a plurality of trained NLP ML models, wherein each ML model of the plurality of trained NLP ML models is trained on a specific target field of target fields; a database to store indexed information; and a memory comprising instructions, coupled to a processor, wherein the instructions are executable by the processor to: input, at least one portion of text into an ML model of the plurality of trained NLP ML models, wherein the ML model is trained on the target field; determine via the ML model, at least one candidate text comprising content related to the target field; extract via the ML model, the at least one candidate text; and generate for each target field of the target fields, via the ML model trained on the target field, an index of the at least one candidate text. ### Claim 14 A system to automatically index multiple informational fields in a digital data record, the system comprising: a plurality of trained NLP ML models, wherein each ML model of the plurality of trained NLP ML models is trained on a specific target field of target fields; a database to store indexed information; and a memory comprising instructions, coupled to a processor, wherein the instructions are executable by the processor to: input, at least one portion of text into an ML model of the plurality of trained NLP ML models, wherein the ML model is trained on the target field; determine via the ML model, at least one candidate text comprising content related to the target field; extract via the ML model, the at least one candidate text; generate for each target field of the target fields, via the ML model trained on the target field, an index of the at least one candidate text; identify, for each target field, based on rules defining each target field, at least one page in the digital data record comprising the content; and extract, for each target field, from the at least one page, at least one portion of text comprising the content. ### Claim 15 A system to automatically index multiple informational fields in a digital data record, the system comprising: a plurality of trained NLP ML models, wherein each ML model of the plurality of trained NLP ML models is trained on a specific target field of target fields; a database to store indexed information; and a memory comprising instructions, coupled to a processor, wherein the instructions are executable by the processor to: input, at least one portion of text into an ML model of the plurality of trained NLP ML models, wherein the ML model is trained on the target field; determine via the ML model, at least one candidate text comprising content related to the target field; extract via the ML model, the at least one candidate text; generate for each target field of the target fields, via the ML model trained on the target field, an index of the at least one candidate text; identify, for each target field, based on rules defining each target field, at least one portion in the digital data record comprising the content; and extract, for each target field, from the at least one portion, at least one portion of text comprising the content. ### Claim 16 (depends on 13) The system of claim 13, wherein the instructions are further executable by the processor to: validate, for each target field, the at least one candidate text. ### Claim 17 (depends on 13) The system of claim 13, wherein the instructions are further executable by the processor to: display the index of the at least one candidate text on a graphical user interface. ### Claim 18 (depends on 13) The system of claim 13, wherein the instructions are further executable by the processor to: combine the index of the at least one candidate text, with at least one other index of an at least one candidate index of at least one other target field of the target fields. ### Claim 19 (depends on 13) The system of claim 13, wherein the instructions are further executable by the processor to: save, for each target field, the index of the at least one candidate text into the database. ### Claim 20 (depends on 13) The system of claim 13, wherein the instructions are further executable by the processor to: identify a page of a digital data record; input an image of the page into a visual machine learning network (visual ML), wherein the visual ML is trained to recognize text associated with the specialized information; identify by the visual ML, a section of the image that contains the specialized information; input a page number of the page, the digital data record, and coordinates of the section into an extraction module; and extract the specialized information by the extraction module from the section. ### Claim 21 (depends on 20) The system of claim 20, wherein the instructions are further executable by the processor to: input the specialized information into a natural language processing ML network (NLP); and identify by the NLP at least one data item based on a presence of the at least one data item in the specialized information and a structure of the specialized information. ### Claim 22 A non-transitory computer-readable storage medium having embodied thereon a program, the program executable by a processor to perform a method for automatically indexing multiple informational target fields in a digital data record comprising: inputting, for each target field, at least one portion of text into a machine learning (ML) model, wherein the ML model is trained on the target field; determining, for each target field, via the ML model, at least one candidate text comprising content related to the target field; and extracting, for each target field, the at least one candidate text. ### Claim 23 A non-transitory computer-readable storage medium having embodied thereon a program, the program executable by a processor to perform a method for automatically indexing multiple informational target fields in a digital data record comprising: inputting, for each target field, at least one portion of text into a machine learning (ML) model, wherein the ML model is trained on the target field; determining, for each target field, via the ML model, at least one candidate text comprising content related to the target field; extracting, for each target field, the at least one candidate text; receiving a digital data record; identifying, for each target field of a plurality of target fields, based on rules defining each target field, at least one page in the digital data record comprising the content; and extracting, for each target field, from the at least one page, at least one portion of text comprising the content. ### Claim 24 (depends on 22) The non-transitory computer-readable storage medium of claim 22, wherein the program executable by the processor further performs: identifying a page of the digital data record; inputting an image of the page into a visual machine learning network (visual ML), wherein the visual ML is trained to recognize text associated with the specialized information; identifying by the visual ML, a section of the image that contains the specialized information; inputting a page number of the page, the digital data record, and coordinates of the section into an extraction module; extracting the specialized information by the extraction module from the section; inputting the specialized information into a natural language processing ML network (NLP); and based on a presence of at least one data item in the specialized information, identifying the at least one data item, by the NLP, based on a structure of the specialized information.