IP Library Granted Patent US 12,620,250
Granted Patent B2
US 12,620,250 · App. 18/084,915 · Granted May 5, 2026

Generalized anomaly detection

Inventors: Olivier Koch (London, GB); Philip Botros (London, GB); Christos Sagonas (London, GB); Francesco Picciotti (London, GB)
Assignee: Onfido Ltd.
G06V30/413G06V10/25G06V10/26
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 12,620,250
App. No.
18/084,915
Granted
May 5, 2026
Kind
B2
Abstract

Described are methods and systems for training a system for detecting anomalies in images of documents in a class of documents. A plurality of training document images of training documents in a class of documents are obtained. For each training document image, the training document image is segmented into a plurality of region of interest (ROI) images, each ROI image corresponding to a respective ROI of the training document. For each ROI image, a plurality of transformations are applied to the ROI image to generate respective transform-specific features for the ROI image and respective transform-specific anomaly scores from the transform-specific features. Based on the respective anomaly scores of the plurality of training document images, a transform-specific threshold is computed for each transformation to separate document images containing an anomaly from document images not containing an anomaly.

Claims (34)

1 . A computer-implemented method of sourcing training images for a system for detecting anomalies in images of documents in a class of documents, the method comprising:

(a) obtaining one or more document images of a fraudulent document;

(b) for each document image:

(i) segmenting the document image into one or more region of interest (ROI) images, each ROI image corresponding to a respective ROI of the fraudulent document;

(ii) generating a vector representation of each ROI image in an embedding space, the embedding space comprising vector representations of one or more other ROI images from a plurality of sample document images;

(iii) determining a similarity between each ROI image from the fraudulent document and one or more other ROI images from sample document images using the vector representations in the embedding space; and

(iv) selecting a plurality of sample document images that have ROI images similar to the ROI images from the fraudulent document and generating a training data set using the selected sample document images; and

(c) in response to a determination that a size of the training data set exceeds a predetermined size threshold, training a supervised anomaly detection machine learning model using the training data set.

2 . The method of claim 1 , wherein the vector representation of each ROI image is generated using a vision transformer (ViT) model.

3 . The method of claim 1 , wherein the similarity is determined using a distance metric.

4 . The method of claim 1 , wherein the one or more document images of the fraudulent document are obtained from an external application programming interface (API) or a user interface of a production anomaly detection system.

5 . The method of claim 1 , wherein at least one of the one or more ROI images comprises an image of the entire fraudulent document.

6 . The method of claim 1 , wherein at least one of the one or more ROI images comprises an image of a portion of the fraudulent document.

7 . The method of claim 1 , wherein generating a vector representation of each ROI image in an embedding space comprises projecting the vector representation into the embedding space.

8 . The method of claim 1 , wherein training a supervised anomaly detection machine learning model using the training data set comprises:

training a plurality of supervised anomaly detection machine learning models using the training data set, each supervised anomaly detection machine learning model corresponding to a respective ROI of the fraudulent document.

9 . A system for sourcing training images for a system for detecting anomalies in images of documents in a class of documents, the system comprising a computing device having a memory for storing computer-executable instructions and a processor that executes the computer-executable instructions to:

(a) obtain one or more document images of a fraudulent document;

(b) for each document image:

(i) segment the document image into one or more region of interest (ROI) images, each ROI image corresponding to a respective ROI of the fraudulent document;

(ii) generate a vector representation of each ROI image in an embedding space, the embedding space comprising vector representations of one or more other ROI images from a plurality of sample document images;

(iii) determine a similarity between each ROI image from the fraudulent document and one or more other ROI images from sample document images using the vector representations in the embedding space; and

(iv) select a plurality of sample document images that have ROI images similar to the ROI images from the fraudulent document and generating a training data set using the selected sample document images;

(c) determine whether a size of the training data set exceeds a predetermined size threshold; and

(d) in response to a determination that the size of the training data set exceeds the predetermined size threshold, train a supervised anomaly detection machine learning model using the training data set; or

(e) in response to a determination that the size of the training data set does not exceed the predetermined size threshold, train an unsupervised anomaly detection machine learning model using the training data set.

10 . The system of claim 9 , wherein the vector representation of each ROI image is generated using a vision transformer (ViT) model.

11 . The system of claim 9 , wherein the similarity is determined using a distance metric.

12 . The system of claim 9 , wherein the one or more document images of the fraudulent document are obtained from an external application programming interface (API) or a user interface of a production anomaly detection system.

13 . The system of claim 9 , wherein at least one of the one or more ROI images comprises an image of the entire fraudulent document.

14 . The system of claim 9 , wherein at least one of the one or more ROI images comprises an image of a portion of the fraudulent document.

15 . The system of claim 9 , wherein generating a vector representation of each ROI image in an embedding space comprises projecting the vector representation into the embedding space.

16 . The system of claim 9 , wherein training a supervised anomaly detection machine learning model using the training data set comprises:

training a plurality of supervised anomaly detection machine learning models using the training data set, each supervised anomaly detection machine learning model corresponding to a respective ROI of the fraudulent document.

Assignments (1)
SECURITY INTEREST Recorded Jul 25, 2024
From: ONFIDO LTD
To: BMO BANK N.A., AS COLLATERAL AGENT
Reel/Frame 068079/0801 →
Priority Claims (2)
EP 21179304 · Jun 14, 2021 · regional
EP 22204513 · Oct 28, 2022 · regional
Continuity (2)
Continuation In Part 17830208 · Jun 1, 2022
Related Publication 20230147685A1 · May 11, 2023
References Cited (59)
US 7590344B2 · Petschnigg · 2009 [cited by applicant]
US 7672475B2 · O'Doherty et al. · 2010 [cited by applicant]
US 8025239B2 · Labrec et al. · 2011 [cited by applicant]
US 8786767B2 · Rihn et al. · 2014 [cited by applicant]
US 9058535B2 · Guigan · 2015 [cited by applicant]
US 9081988B2 · Dolev · 2015 [cited by applicant]
US 9153005B2 · Tremolada et al. · 2015 [cited by applicant]
US 9418282B2 · Rutz et al. · 2016 [cited by applicant]
US 10019626B2 · Schilling et al. · 2018 [cited by applicant]
US 10019627B2 · Kutter et al. · 2018 [cited by applicant]
US 10043071B1 · Wu · 2018 [cited by applicant]
US 10127454B2 · Pau et al. · 2018 [cited by applicant]
US 10354142B2 · Arlazarov et al. · 2019 [cited by applicant]
US 10402448B2 · Filgueiras De Araujo et al. · 2019 [cited by applicant]
US 10477087B2 · Rivard et al. · 2019 [cited by applicant]
US 10621426B2 · Gaubatz et al. · 2020 [cited by applicant]
US 10789463B2 · Kutter et al. · 2020 [cited by applicant]
US 10872488B2 · Azanza Ladrn et al. · 2020 [cited by applicant]
US 11354917B2 · Martin · 2022 [cited by applicant]
US 11450152B2 · Ramachandra · 2022 [cited by applicant]
US 20060281435A1 · Shearer · 2006 [cited by applicant]
US 20070041628A1 · Fan · 2007 [cited by applicant]
US 20080158258A1 · Lazarus et al. · 2008 [cited by applicant]
US 20140037196A1 · Blair · 2014 [cited by applicant]
US 20140055824A1 · Tremolada et al. · 2014 [cited by applicant]
US 20170132465A1 · Kutter et al. · 2017 [cited by applicant]
US 20190220660A1 · Cali et al. · 2019 [cited by applicant]
US 20190392196A1 · Sagonas et al. · 2019 [cited by applicant]
US 20200184212A1 · Anthony Samy · 2020 [cited by examiner]
US 20200311409A1 · Luo et al. · 2020 [cited by applicant]
US 20200379868A1 · Dherange · 2020 [cited by applicant]
US 20210004949A1 · Broyda · 2021 [cited by applicant]
US 20210019519A1 · Martin et al. · 2021 [cited by applicant]
US 20210124919A1 · Balakrishnan et al. · 2021 [cited by applicant]
US 20210160687A1 · Ross et al. · 2021 [cited by applicant]
US 20210200855A1 · Guigan · 2021 [cited by applicant]
US 20220284301A1 · Qiu et al. · 2022 [cited by applicant]
US 20220351532A1 · Gietema · 2022 [cited by applicant]
US 20230069960A1 · Mahadevan et al. · 2023 [cited by applicant]
US 20230281959A1 · Hoshen · 2023 [cited by examiner]
US 20240242487A1 · Conde · 2024 [cited by examiner]
DE 102018102015A1 · 2019 [cited by applicant]
EP 4105825A1 · 2022 [cited by applicant]
EP 4361971A1 · 2024 [cited by applicant]
FR 2977349A3 · 2013 [cited by applicant]
WO 2020255137A1 · 2020 [cited by applicant]
WO 2021044082A1 · 2021 [cited by applicant]
Hudson et al “Generative Adversarial Transformers” Cornell University Library, Mar. 2, 2021. [cited by applicant]
“CARS 2020—Computer Assisted Radiology and Surgery Proceedings of the 34th International Congress and Exhibition, Munich, Germany, Jun. 23-27, 2020 ED—Barratt D; Jannin P; Fitchinger G; Cotin S,” International Journal o… [cited by applicant]
Dosovitskiy, Alexey et al., “An image is worth 16×16 words: Transformers for image recognition at scale,” Jun. 3, 2021, pp. 1-22. [cited by applicant]
G. Evangelidis and E. Psarakis, “Parametric Image Alignment Using Enhanced Correlation Coefficient Maximization,” IEEE Transactions on Pattern Analysis and Machine Intelligence 30(10), Oct. 2008, 1858-1865. [cited by applicant]
N. Ahmed, T. Natarajan, and K.R. Rao, “Discrete Cosine Transform,” IEEE Transactions on Computers, vol. C-23, Issue 1, Jan. 1974, pp. 90-93. [cited by applicant]
H. Abdi and L.J. Williams, “Principal Component Analysis,” Wiley Interdisciplinary Reviews: Computational Statistics, vol. 2, Issue 4, Jul. 15, 2010, pp. 433-459. [cited by applicant]
R. Haralick, K. Shanmugam, and I.H. Dinstein, “Textural Features for Image Classification,” IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-3, Issue 6, Nov. 1973, pp. 610-621. [cited by applicant]
T. Ojala, M. Pietikainen, and T. Maenpaa, “Multiresolution Gray-scale and Rotation Invariant Texture Classification with Local Binary Patterns,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, I… [cited by applicant]
D. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” International Journal of Computer Vision, 60(2), Nov. 2004, pp. 91-110. [cited by applicant]
K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv:1409.1556 [cs.CV], Apr. 10, 2015. [cited by applicant]
Search Report issued in EP4105825, completed Nov. 15, 2021. [cited by applicant]
Search Report issued in EP4361971, completed Mar. 17, 2023. [cited by applicant]