US 12,260,174 B2
Detection of altered documents
Ralph Mayer, Encinitas, CA (US); and Erik Giles, Encinitas, CA (US)
Assigned to MONEYTHUMB, INC., Encinitas, CA (US)
Appl. No. 18/027,855
Filed by MoneyThumb LLC, Encinitas, CA (US)
PCT Filed Sep. 21, 2021, PCT No. PCT/US2021/051363
§ 371(c)(1), (2) Date Mar. 22, 2023,
PCT Pub. No. WO2022/066666, PCT Pub. Date Mar. 31, 2022.
Claims priority of provisional application 63/081,453, filed on Sep. 22, 2020.
Prior Publication US 2023/0359815 A1, Nov. 9, 2023
Int. Cl. H04L 9/00 (2022.01); G06F 40/194 (2020.01)
CPC G06F 40/194 (2020.01) 17 Claims
OG exemplary drawing
 
1. A method of determining third party document authenticity comprising using at least one hardware processor to:
access a single instance of a portable document format (PDF) document for a first time, the PDF document generated by an unknown party, wherein the PDF document is asserted to be generated by a particular third party;
extract document content information from the PDF document comprising a plurality of PDF data objects and a plurality of PDF commands, wherein the plurality of PDF data objects include at least one text string and the plurality of PDF commands include at least one PDF command configured to position a text string;
analyze the extracted document content information and generate an intra document model for the PDF document comprising one or more combinations of PDF command and PDF data object used to position a text string;
evaluate the document content information in accordance with the intra document model;
identify one or more first artifacts in the document content information based on the evaluation in accordance with the intra document model, wherein the one or more first artifacts includes the presence of a first combination of PDF operator and PDF data object used to position a first text string in a set of text strings and the presence of a second combination of PDF command and PDF data object used to position a second text string in the same set of text strings;
determine an intra document score for the PDF document based on the identified one or more first artifacts;
evaluate the document content information in accordance with one or more inter document models, wherein an inter document model comprises a set of consistencies across a plurality of known PDF documents and includes a set of known combinations of PDF operator and PDF data object used by the particular third party to position a text string;
identify one or more second artifacts in the document content information based on the evaluation in accordance with the one or more inter document models, wherein the one or more second artifacts includes the presence of a second combination of PDF operator and PDF data object used to position a second text string not found in the set of known combinations of PDF operator and PDF data object used by the particular third party to position a text string included in the inter document model;
determine an inter document score for the PDF document based on the identified one or more second artifacts; and
determine an alteration score for the PDF document based on the intra document score and the inter document score, wherein a fraud detection level is determined as high risk when the alteration score exceeds a predetermined threshold.