IP Library › Granted Patent US 12,743,759
Granted Patent B2
US 12,743,759 · App. 18/401,422 · Granted Sep 22, 2026

Multiple fraud type detection system and methods

Inventors: Daryl Huff (Saratoga, CA); Lei Guang (Montreal, CA); Paras Kapoor (Montreal, CA); Hasmik Martirosyan (Ontario, CA); Alix Melchy (Montreal, CA); Artem Voronin (Montreal, CA); Stuart Wells (Sunnyvale, CA)
Assignee: Jumio Corporation
G06T7/0002G06T7/194G06V40/172
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,743,759
App. No.
18/401,422
Granted
Sep 22, 2026
Kind
B2
Abstract

A system and method for multiple fraud type detection includes anti-injection attack system that has a layered architectural approach that uses a includes combination of different specific models to detect the attacks in combination with image processing techniques, device signals and liveness checks to detect the variety of different types of fraud attacks or repeat fraud attacks. The anti-injection attack system applies the analysis of the tools used to create deepfake, face morph and face swap attacks to define the elements of its layered architecture that can these various types of attacks.

Claims (58)

1 . A computer-implemented method comprising:

receiving first image data associated with a user;

performing a first detection type comprising a multi-frame deepfake detection model on the first image data to generate a first signal;

performing a second detection type comprising a single frame deepfake detection model on the first image data to generate a second signal;

performing a third detection type comprising a subject and scene segmentation analysis on the first image data, wherein performing the third detection type further comprises generating a vector embedding for the first image data, accessing a matrix of vector embeddings, searching for cosine similarities between embeddings in the matrix and the vector embedding generated, and generating a third signal based on the searching for the cosine similarities;

generating an aggregated fraud score based upon the first signal, the second signal and the third signal; and

accepting the first image data as genuine upon comparing the aggregated fraud score to a threshold to determine if the aggregated fraud score satisfies the threshold.

2 . The computer-implemented method of claim 1 further comprising: rejecting the first image data as fraudulent based upon the aggregated fraud score.

3 . The computer-implemented method of claim 1 wherein the first detection type is one from a group of: a deepfake model, a face morph model, a face swap model, an unknown injection attack anomaly model, a subject and scene segmentation analyzer, an injection checker, a device risk checker, a liveness detector, a face match and face analysis subsystem, an injection attack detector and a device risk detector.

4 . The computer-implemented method of claim 1 wherein the second detection type is one from a group of: a deepfake model, a face morph model, a face swap model, an unknown injection attack anomaly model, a subject and scene segmentation analyzer, an injection checker, a device risk checker, a liveness detector, a face match and face analysis subsystem, an injection attack detector and a device risk detector.

5 . The computer-implemented method of claim 1 wherein the first image data includes a selfie image and a document image, and the method further comprises performing a fourth detection type using the selfie image and the document image to generate a fourth signal, wherein the fourth detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the fourth signal based on a match between the selfie image and the holder portion of the document image and

wherein the aggregated fraud score is generated based upon the first signal, the second signal, the third signal, and the fourth signal.

6 . The computer-implemented method of claim 1 wherein the first image data includes a selfie image and a document image, and the method further comprises performing the third detection type using the selfie image and the document image to generate the third signal, wherein the third detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the third signal based on a match between the selfie image and the holder portion of the document image, and wherein the first image data is rejected as genuine based upon only the third signal.

7 . The computer-implemented method of claim 1 wherein the first image data is one from a group of a selfie and a video.

8 . The computer-implemented method of claim 1 wherein the first image data includes a selfie image, and the method further comprises:

receiving a document image;

performing the first detection type on the document image to generate the third signal;

performing the second detection type on the document image to generate a fourth signal; and

wherein the generating the aggregated fraud score is also based on the third signal and the fourth signal.

9 . The computer-implemented method of claim 1 wherein the first image data includes a selfie image, and the method further comprises:

receiving a document image;

extracting selfie metadata from the selfie image;

extracting document metadata from the document image; and

wherein the generating the aggregated fraud score is also based on the selfie metadata and the document metadata.

10 . The computer-implemented method of claim 1 wherein the first image data includes a selfie image and a document image, and the method further comprises performing the third detection type using the selfie image and the document image to generate the third signal, wherein the third detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the third signal based on a match between the selfie image and the holder portion of the document image.

11 . The computer-implemented method of claim 1 wherein satisfying the threshold indicates that the first image data is acceptable and not satisfying the threshold indicates that the first image data is fraudulent.

12 . The computer-implemented method of claim 1 , further comprising:

performing the third detection type on the first image data to generate the third signal; and

wherein the first image data is rejected as genuine based upon only the third signal.

13 . A system comprising:

one or more processors; and

a memory operably coupled with the one or more processors, wherein the memory stores instructions that, in response to execution of the instructions by the one or more processors, cause the one or more processors to:

receive a first image data associated with a user;

perform a first detection type comprising a multi-frame deepfake detection model on the first image data to generate a first signal;

perform a second detection type comprising a single frame deepfake detection model on the first image data to generate a second signal;

perform a third detection type comprising a subject and scene segmentation analysis on the first image data, wherein performing the third detection type further comprises generating a vector embedding for the first image data, accessing a matrix of vector embeddings, searching for cosine similarities between embeddings in the matrix and the vector embedding generated, and generating a third signal based on the searching for the cosine similarities;

generate an aggregated fraud score based upon the first signal and the second signal and the third signal; and

accept the first image data as genuine based upon comparing the aggregated fraud score to a threshold to determine if the aggregated fraud score satisfies the threshold.

14 . The system of claim 13 , wherein the instructions cause the one or more processors to reject the first image data as fraudulent based upon the aggregated fraud score.

15 . The system of claim 13 , wherein the first detection type is one from a group of: a deepfake model, face morph model, a face swap model, an unknown injection attack anomaly model, a subject and scene segmentation analyzer, an injection checker, a device risk checker, a liveness detector, a face match and face analysis subsystem, an injection attack detector, and a device risk detector.

16 . The system of claim 13 , wherein the second detection type is one from a group of: a deepfake model, face morph model, a face swap model, an unknown injection attack anomaly model, a subject and scene segmentation analyzer, an injection checker, a device risk checker, a liveness detector, a face match and face analysis subsystem, an injection attack detector and a device risk detector.

17 . The system of claim 13 , wherein the first image data includes a selfie image and a document image, and the system is further configured to performing a fourth detection type using the selfie image and the document image to generate a fourth signal, wherein the fourth detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the fourth signal based on a match between the selfie image and the holder portion of the document image, and wherein the aggregated fraud score is generated based upon the first signal, the second signal, the third signal, and the fourth signal.

18 . The system of claim 13 14 , wherein the first image data includes a selfie image and a document image, and the system is further configured to performing the third detection type using the selfie image and the document image to generate the third signal, wherein the third detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the third signal based on a match between the selfie image and the holder portion of the document image, and wherein the first image data is rejected as genuine based upon only the third signal.

19 . The system of claim 13 , wherein the first image data is one from a group of a selfie and a video.

20 . The system of claim 13 , wherein the first image data includes a selfie image, and the instructions cause the one or more processors to:

receive a document image;

perform the first detection type on the document image to generate the third signal;

perform the second detection type on the document image to generate a fourth signal; and

wherein the aggregated fraud score is also based on the third signal and the fourth signal.

21 . The system of claim 13 , wherein the first image data includes a selfie image, and the instructions cause the one or more processors to:

receive a document image;

extract selfie metadata from the selfie image; and

extract document metadata from the document image, and wherein the aggregated fraud score is also based on the selfie metadata and the document metadata.

22 . The system of claim 13 wherein the first image data includes a selfie image and a document image, and wherein the instructions cause the one or more processors to perform the third detection type using the selfie image and the document image to generate the third signal, wherein the third detection type is a face match and face analysis that compares the selfie image to a holder portion of the document image and generates the third signal based on a match between the selfie image and the holder portion of the document image.

23 . The system of claim 13 wherein satisfying the threshold indicates that the first image data is acceptable and not satisfying the threshold indicates that the image data is fraudulent.

24 . The system of claim 13 , wherein the instructions cause the one or more processors to:

perform the third detection type on the first image data to generate the third signal; and

wherein the first image data is rejected as genuine based upon only the third signal.

Assignments (2)
CORRECTIVE ASSIGNMENT TO CORRECT THE LIST OF INVENTORS TO INCLUDE STUART WELLS PREVIOUSLY RECORDED ON REEL 66085 FRAME 608. ASSIGNOR(S) HEREBY CONFIRMS THE ASSIGNMENT. Recorded Jan 17, 2024
From: HUFF, DARYL; GUANG, LEI; KAPOOR, PARAS; MARTIROSYAN, HASMIK; MELCHY, ALIX; VORONIN, ARTEM; WELLS, STUART
To: JUMIO CORPORATION
Reel/Frame 066551/0075 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 10, 2024
From: HUFF, DARYL; GUANG, LEI; KAPOOR, PARAS; MARTIROSYAN, HASMIK; MELCHY, ALIX; VORONIN, ARTEM
To: JUMIO CORPORATION
Reel/Frame 066085/0608 →
Continuity (1)
Related Publication 20250217952A1 · Jul 3, 2025
References Cited (169)
US 4897814A · Clark · 1990 [cited by applicant]
US 5425110A · Spitz · 1995 [cited by applicant]
US 5544255A · Smithies et al. · 1996 [cited by applicant]
US 5668897A · Stolfo · 1997 [cited by applicant]
US 5748780A · Stolfo · 1998 [cited by applicant]
US 5838814A · Moore · 1998 [cited by applicant]
US 6363162B1 · Moed et al. · 2002 [cited by applicant]
US 7040539B1 · Stover · 2006 [cited by applicant]
US 7831531B1 · Baluja et al. · 2010 [cited by applicant]
US 8326761B1 · Hecht et al. · 2012 [cited by applicant]
US 8352494B1 · Badoiu · 2013 [cited by applicant]
US 8886648B1 · Procopio et al. · 2014 [cited by applicant]
US 8910032B2 · Graves et al. · 2014 [cited by applicant]
US 9135517B1 · Adams · 2015 [cited by applicant]
US 9836591B2 · John Archibald · 2017 [cited by examiner]
US 10320807B2 · Khan · 2019 [cited by applicant]
US 10628702B1 · Gerstner et al. · 2020 [cited by applicant]
US 11144752B1 · Castelblanco et al. · 2021 [cited by applicant]
US 11416562B1 · Gruhl et al. · 2022 [cited by applicant]
US 11593439B1 · Avadhani et al. · 2023 [cited by applicant]
US 11625954B2 · Kwak · 2023 [cited by examiner]
US 11715102B2 · Edwards et al. · 2023 [cited by applicant]
US 11900755B1 · Bueche, Jr. · 2024 [cited by applicant]
US 12197483B1 · Shmukler et al. · 2025 [cited by applicant]
US 12387512B1 · Sureka et al. · 2025 [cited by applicant]
US 20020064305A1 · Taylor · 2002 [cited by applicant]
US 20030172066A1 · Cooper et al. · 2003 [cited by applicant]
US 20030229637A1 · Baxter et al. · 2003 [cited by applicant]
US 20040155877A1 · Hong et al. · 2004 [cited by applicant]
US 20060026156A1 · Zuleba · 2006 [cited by applicant]
US 20060041506A1 · Mason et al. · 2006 [cited by applicant]
US 20060124726A1 · Kotovich et al. · 2006 [cited by applicant]
US 20060164682A1 · Lev · 2006 [cited by applicant]
US 20070078846A1 · Gulli et al. · 2007 [cited by applicant]
US 20070086628A1 · Fuchs et al. · 2007 [cited by applicant]
US 20070116328A1 · Sablak · 2007 [cited by applicant]
US 20070150387A1 · Seubert et al. · 2007 [cited by applicant]
US 20070256010A1 · Blackmon et al. · 2007 [cited by applicant]
US 20080065630A1 · Luo et al. · 2008 [cited by applicant]
US 20080149713A1 · Brundage · 2008 [cited by applicant]
US 20090060396A1 · Blessan et al. · 2009 [cited by applicant]
US 20090152357A1 · Lei et al. · 2009 [cited by applicant]
US 20090261158A1 · Lawson · 2009 [cited by applicant]
US 20100027896A1 · Geva et al. · 2010 [cited by applicant]
US 20110057040A1 · Jones · 2011 [cited by applicant]
US 20110128360A1 · Hatzav et al. · 2011 [cited by applicant]
US 20130129159A1 · Huijgens et al. · 2013 [cited by applicant]
US 20140002872A1 · Cook · 2014 [cited by applicant]
US 20140046954A1 · Maclean et al. · 2014 [cited by applicant]
US 20140161326A1 · Ganong et al. · 2014 [cited by applicant]
US 20150100590A1 · Robinson et al. · 2015 [cited by applicant]
US 20150242592A1 · Weiss et al. · 2015 [cited by applicant]
US 20150341370A1 · Khan · 2015 [cited by applicant]
US 20160005050A1 · Teman · 2016 [cited by applicant]
US 20160098399A1 · Casperson · 2016 [cited by applicant]
US 20160210450A1 · Su · 2016 [cited by applicant]
US 20170161375A1 · Stoica et al. · 2017 [cited by applicant]
US 20170193285A1 · Negi · 2017 [cited by applicant]
US 20170277945A1 · Budihal et al. · 2017 [cited by applicant]
US 20170322932A1 · Deschenes et al. · 2017 [cited by applicant]
US 20170337449A1 · Hamada et al. · 2017 [cited by applicant]
US 20170351909A1 · Kaehler · 2017 [cited by applicant]
US 20180060874A1 · Kelts et al. · 2018 [cited by applicant]
US 20180075090A1 · Knight et al. · 2018 [cited by applicant]
US 20180186164A1 · Wu · 2018 [cited by applicant]
US 20180204113A1 · Galron et al. · 2018 [cited by applicant]
US 20180293461A1 · Le et al. · 2018 [cited by applicant]
US 20180300296A1 · Ziraknejad et al. · 2018 [cited by applicant]
US 20180373859A1 · Ganong · 2018 [cited by applicant]
US 20190035431A1 · Attorre · 2019 [cited by applicant]
US 20190205686A1 · Mayer et al. · 2019 [cited by applicant]
US 20190272549A1 · Mossoba et al. · 2019 [cited by applicant]
US 20190278986A1 · Nepomniachtchi · 2019 [cited by applicant]
US 20200184201A1 · Kaehler · 2020 [cited by applicant]
US 20200304650A1 · Roach · 2020 [cited by applicant]
US 20200342600A1 · Sjstrand et al. · 2020 [cited by applicant]
US 20200366671A1 · Larson et al. · 2020 [cited by applicant]
US 20210075788A1 · Pasterk et al. · 2021 [cited by applicant]
US 20210124919A1 · Balakrishnan · 2021 [cited by applicant]
US 20210174016A1 · Fox et al. · 2021 [cited by applicant]
US 20210248401A1 · Timoshenko · 2021 [cited by examiner]
US 20210259660A1 · Bharat et al. · 2021 [cited by applicant]
US 20210272341A1 · Swaminathan et al. · 2021 [cited by applicant]
US 20210307841A1 · Buch et al. · 2021 [cited by applicant]
US 20210320801A1 · Wyss · 2021 [cited by applicant]
US 20210326461A1 · Paul et al. · 2021 [cited by applicant]
US 20210326629A1 · Slattery · 2021 [cited by applicant]
US 20220028086A1 · Woodard et al. · 2022 [cited by applicant]
US 20220058660A1 · Ivanov · 2022 [cited by applicant]
US 20220114456A1 · Nouri et al. · 2022 [cited by applicant]
US 20220180113A1 · Patel et al. · 2022 [cited by applicant]
US 20220182430A1 · Bennett-James et al. · 2022 [cited by applicant]
US 20220385880A1 · Nims · 2022 [cited by applicant]
US 20230013380A1 · Choi et al. · 2023 [cited by applicant]
US 20230017185A1 · Cheong et al. · 2023 [cited by applicant]
US 20230083000A1 · Fujimoto et al. · 2023 [cited by applicant]
US 20230113148A1 · Zlotnick · 2023 [cited by applicant]
US 20230129350A1 · Bryan et al. · 2023 [cited by applicant]
US 20230143239A1 · Yusuf et al. · 2023 [cited by applicant]
US 20230196628A1 · Bischoff et al. · 2023 [cited by applicant]
US 20230298031A1 · Drapeau et al. · 2023 [cited by applicant]
US 20230401824A1 · Khan et al. · 2023 [cited by applicant]
US 20230421602A1 · Boyer et al. · 2023 [cited by applicant]
US 20240046686A1 · Ye et al. · 2024 [cited by applicant]
US 20240193970A1 · Hsu et al. · 2024 [cited by applicant]
US 20240202294A1 · Yogerst et al. · 2024 [cited by applicant]
US 20240205239A1 · Bonev · 2024 [cited by examiner]
US 20240411982A1 · Malanga et al. · 2024 [cited by applicant]
US 20250005950A1 · Eren · 2025 [cited by applicant]
US 20250218226A1 · Marshalkin · 2025 [cited by applicant]
US 20250225527A1 · Jiang et al. · 2025 [cited by applicant]
AU 2018100581A4 · 2018 [cited by applicant]
MY 192715A · 2022 [cited by applicant]
WO 2022015948A1 · 2022 [cited by applicant]
Di Guardo, Fabrizio. “Facemask-a Real-Time Face Morphing Tool.” Medium, Level Up Coding, May 2, 2023, levelup.gitconnected.com/facemask-a-real-time-face-morphing-tool-5b343591a237. May 2, 2023. [cited by applicant]
Kirillov, Alexander et al. “Segment Anything.” 2023 IEEE/CVF International Conference on Computer Vision (ICCV), Oct. 1, 2023, doi:10.1109/iccv51070.2023.00371. Oct. 1, 2023. [cited by applicant]
Liu, Haotian, et al. “Visual Instruction Tuning.” 37th Conference on Neural Information Processing Systems, Dec. 11, 2023, arxiv.org/pdf/2304.08485. Dec. 11, 2023. [cited by applicant]
Wang, Xin, et al. “Attribute-Aware Implicit Modality Alignment for Text Attribute Person Search.” Arxiv.Org, Jun. 6, 2024. Jun. 6, 2024. [cited by applicant]
Wang, Zhe, et al. “Attribute-guided Transformer for Robust Person Re-identification.” IET Computer Vision, vol. 17, No. 8, Jun. 23, 2023, pp. 977-992, doi:10.1049/cvi2.12215. Jun. 23, 2023. [cited by applicant]
“7 Best Face Morph Apps 2024 (Morph Two Faces Together).” ContentMaverickscom, Content Mavericks, 2024, contentmavericks.com/best-face-morph-app/. 2024 Web. 25 pgs. 2024. [cited by applicant]
“Levenshtein Distance.” Wikipedia, Wikimedia Foundation, Dec. 23, 2023, en.wikipedia.org/wiki/Levenshtein_distance. Web. 6 pgs. 2023. [cited by applicant]
“New Method Detects Deepfake Videos with up to 99% Accuracy.” News, May 3, 2022, news.ucr.edu/articles/2022/05/03/new-method-detects-deepfake-videos-99-accuracy. Web. 3 pgs. 2022. [cited by applicant]
Aslam, Asra, et al., May 15, 2019, Depth-Map Generation using Pixel Matching in Stereoscopic Pair of Images, https://arxiv.org/pdf/1902.03471.pdf, 5 pgs. 2019. [cited by applicant]
Bassil, Youssef, and Mohammad Alwani. “Context-Sensitive Spelling Correction Using Google Web 1T 5-Gram Information.” Computer and Information Science (Toronto), vol. 5, No. 3, 2012, p. 37. 2012. [cited by applicant]
Bassil, Youssef, and Mohammad Alwani. “OCR Post-Processing Error Correction Algorithm Using Google Online Spelling Suggestion.” ArXiv.org, 2012, pp. arXiv.org, 2012. 2012. [cited by applicant]
Benalcazar, Daniel, et al. “Synthetic ID Card Image Generation for Improving Presentation Attack Detection.” IEEE Transactions on Information Forensics and Security, vol. 18, 2023, pp. 1814-1824. 2023. [cited by applicant]
Canada Passport Phot Security Features, Canada.ca, Government of Canada (Dec. 20, 2022) https://www.canada.ca/en/immigration-refugees-citizenship/services/canadian-passports/photos.html#photo, webpage 11 pgs. 2022. [cited by applicant]
Casado, Constantino Alvarez, et al. “Real-time Face Alignment: Evaluation Methods, Training Strategies and Implementation Optimization.” Journal of Real-time Image Processing 18.6 (2021): 2239-2267. Web. 2021. [cited by applicant]
deepswap.ai, “Deepswap—Best Face and Video Edit Tools Online.” DeepSwap AI, www.deepswap.ai/?utm_source=bing&cp_id=441169896&msclkid=eb76125c4d531cf9797e77325a694067. Accessed Dec. 2023. Web. 11 pgs. 2023. [cited by applicant]
DevCodeF1 Editors, Using LLMS for OCR text proofreading: A guide for software developers. Dev Code F1. May 26, 2023, <https://devcodef1.com/news/1007434/llms-for-ocr-text-proofreading> Web. 3 pgs. 2023. [cited by applicant]
Edge Detection Using OpenCV, LearnOpenCV.com, (2023) https://learnopencv.com/edge-detection-using-opencv/, webpage. 10 pgs. 2023. [cited by applicant]
Elsayed, M et al. “A New Method for Full Reference Image Blur Measure.” International Journal of Simulation: Systems Science and Technology V19 N1 (Feb. 1, 2018): 7.1-7.5 2018 https://doi.org/10.5013/IJSSST.a.19.01.7. 2… [cited by applicant]
Faceshape. “Face Morphing Simulator—Morph Two Faces Together.” Face Morphing Simulator—Morph Two Faces Together, www.faceshape.com/face-morph. (2022) Web. 2 pgs. 2022. [cited by applicant]
Fadilpašić, Sead. “Deepfake Fraud Attacks Are Hitting More and More Businesses.” TechRadar, TechRadar Pro, Feb. 24, 2023, www.techradar.com/news/deepfake-fraud-attacks-are-hitting-more-and-more-businesses. Accessed Jan.… [cited by applicant]
GitHub, Use Llama2 to Improve the Accuracy of Tesseract OCR. GitHub. (n.d.). https://github.com/Dicklesworthstone/llama2_aided_tesseract> Web. Last updated Aug. 2, 2023. 3 pgs. 2023. [cited by applicant]
Help Net Security. “Detecting Face Morphing: A Simple Guide to Countering Complex Identity Fraud.” Help Net Security, Mar. 16, 2023, www.helpnetsecurity.com/2023/03/20/facial-morphing-technology/. Web. 5 pgs. 2023. [cited by applicant]
Hu, Yifei, et al. “Misspelling Correction with Pre-Trained Contextual Language Model.” ArXiv.org, 2021, pp. arXiv.org, 2021. 2021. [cited by applicant]
Kaskpersky. “What Does the Rise of Deepfakes Means for the Future of Cybersecurity?” Daily English USA Usakasperskycomblog, Jul. 16, 2023, USA.kaspersky.com/blog/secure-futures-magazine/deepfakes-2019/21932/. [cited by applicant]
Kramer, Robin S. S., et al. “Face Morphing Attacks: Investigating Detection with Humans and Computers.” Cognitive Research: Principles and Implications, vol. 4, No. 1, 2019, p. 28. 2019. [cited by applicant]
Kumar, Varun. “14 Best Deepfake Apps and Tools in 2024.” RankRed, Jan. 1, 2024, www.rankred.com/best-deepfake-apps-tools/. Web. 22 pgs. 2024. [cited by applicant]
Lee, Jung-Hun, et al. “Deep Learning-Based Context-Sensitive Spelling Typing Error Correction.” IEEE Access, vol. 8, 2020, pp. 152565-152578. 2020. [cited by applicant]
Liu, Zhaoxiang, et al. “Facial Pose Estimation by Deep Learning from Label Distributions.” (2019). Web. IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), 9 pgs. 2019. [cited by applicant]
Lowphansirkul, Lalita, et al. “WangchanBERTa: Pretraining Transformer-Based Thai Language Models.” ArXiv.org, 2021, pp. arXiv.org, 2021. 2021. [cited by applicant]
Luo, Yinhui et al. “A Review of Homography Estimation: Advances and Challenges.” Electronics 2023 pp. 4977-4977. https://www.mdpi.com/2079-9292/12/24/4977 2023. [cited by applicant]
Lyu, Siwei, “Detecting ‘deepfake’ Videos in the Blink of an Eye.” The Conversation, Sep. 15, 2022, theconversation.com/detecting-deepfake-videos-in-the-blink-of-an-eye-101072. Web. 4 pgs. 2022. [cited by applicant]
Mazaheri, Ghazal et al. “2022 Ieee/Cvf Winter Conference on Applications of Computer Vision (Wacv).” Detection and Localization of Facial Expression Manipulations IEEE 2022 pp. 2773-2783 2022. [cited by applicant]
Nightingale, Sophie J., et al. “Perceptual and Computational Detection of Face Morphing.” Journal of Vision (Charlottesville, Va.), vol. 21, No. 3, 2021, p. 4. 2021. [cited by applicant]
Passport Photo Specifications., Government of Canada (2015) https://www.canada.ca/content/dam/ircc/migration/ircc/english/pdf/pub/pass-photo-spec-eng.pdf, 5 pgs. 2015. [cited by applicant]
picsi.ai, “Create Realistic Face Morphs” Picsi.AI, InsightFace, 2023, www.picsi.ai/. Web. 8 pgs. 2023. [cited by applicant]
Rosebrock, A. (Nov. 16, 2021). OCR passports with opencv and Tesseract. PylmageSearch. <https://pyimagesearch.com/2021/12/01/ocr-passports-with-opencv-and-tesseract/> Web. 24 pgs. 2021. [cited by applicant]
Rosebrock, Adrian, “OpenCV Fast Fourier Transform (FFT) for blur detection in images and video streams.” PyImageSearch.com, Jun. 15, 2020, webpage, 21 pgs. https://pyimagesearch.com/2020/06/15/opencv-fast-fourier-transf… [cited by applicant]
Rosebrock, Adrian, Blur Detection with OpenCV PyImageSearch.com, Sep. 7, 2015, https://pyimagesearch.com/2015/09/07/blur-detection-with-opencv/, webpage, 13 pgs. 2015. [cited by applicant]
Sadeghzadeh, Arezoo, et al., “Pose-invariant face recognition based on matching the occlusion free regions aligned by 3D generic model.” IET Computer Vision, Aug. 2020, vol. 14, Issue 5, pp. 177-287. 2020. [cited by applicant]
Schulz, Daniel, et al., “Identify Documents Image Quality Assessment.” European Association for Signal Processing, Proceedings 2022 pp. 1017-1021. 2022. [cited by applicant]
Seibold, C., et al., “Detection of Face Morphing Attacks by Deep Learning.” Digital Forensics and Watermarking. IWDW (2017). Lecture Notes in Computer Science(), vol. 10431. Springer, Cham. https://doi.org/10.1007/978-3… [cited by applicant]
Shokat, Sana, et al. “Analysis and Evaluation of Braille to Text Conversion Methods.” Mobile Information Systems, vol. 2020, 2020, pp. 1-14. 2020. [cited by applicant]
Shokat, Sana, et al. “Characterization of English Braille Patterns Using Automated Tools and RICA Based Feature Extraction Methods.” Sensors (Basel, Switzerland), vol. 22, No. 5, 2022, p. 1836. 2022. [cited by applicant]
Tesseract—OCR, Improving the Quality of the Output Github.com, Tesseract User Manual v. 5.x, 9 pgs, updated Dec. 5, 2023. Web. 2023. [cited by applicant]
Times, Global. “Tencent Launches Large Language Model ‘Hunyuan’ amid Global Generative Ai Frenzy.” Global Times, Sep. 7, 2023, www.globaltimes.cn/page/202309/1297761.shtml. 6 pgs. 2023. [cited by applicant]
YouTube, YouTube, “Morphing Identity: A real-time face morphing system to transforming face identity.” May 15, 2021, Cybernetic Humanity Studio. https://youtu.be/ahKxwaJ3k_U?si=kp4G4wFa8IAsGNLY 2021. [cited by applicant]
Zhang, Erhu, et al. “Forgery Detection for Perforated Number in Security Document by Analysing the Perforated Holes.” The Imaging Science Journal 65.1 (2017): 40-48. Web. 2017. [cited by applicant]
Zhao, Jian et al. “2018 Ieee/Cvf Conference on Computer Vision and Pattern Recognition.” Towards Pose Invariant Face Recognition in the Wild IEEE 2018 pp. 2207-2216. 2018. [cited by applicant]
PCT International Search Report and Written Opinion; Application No. PCT/US24/62269 Jumio Corporation, International filing date of Dec. 30, 2024, date of mailing Mar. 7, 2025, 11 pages. [cited by applicant]
PCT International Search Report and Written Opinion; Application No. PCT/US2024/048525 Jumio Corporation, International filing date of Sep. 26, 2024, date of mailing Nov. 27, 2024, 10 pages. 2024. [cited by applicant]
PCT International Search Report and Written Opinion; Application No. PCT/US23/79821 Jumio Corporation, International filing date of Nov. 15, 2023, date of mailing Apr. 4, 2024, 10 pages. [cited by applicant]
PCT International Search Report and Written Opinion; Application No. PCT/US23/86219 Jumio Corporation, International filing date of Dec. 28, 2023, date of mailing May 23, 2024, 4 pages. [cited by applicant]
PCT International Search Report and Written Opinion; Application No. PCT/US23/79511 Jumio Corporation, International filing date of Nov. 13, 2023, date of mailing Mar. 4, 2024, 17 pages. [cited by applicant]
Huang, Jing, et al. “A multiplexed network for end-to-end, multilingual OCR.” Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2021. (Year: 2021). [cited by applicant]
Nguyen, Thi Tuyet Hai, et al. “Neural machine translation with BERT forpost-OCR error detection and correction.” Proceedings of the ACM/IEEE joint conference on digital libraries in 2020. 2020. (Year: 2020). [cited by applicant]