# US 11,308,492 B2 · Anomaly and fraud detection with fake event detection using pixel intensity testing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,308,492 B2
- Document ID: US11308492B2
- Status: Granted
- Kind: B2
- Application number: 16/711,679
- Filed: 2019-12-12
- Published: 2021-01-07
- Granted: 2022-04-19
- Earliest claimed: 2019-07-03
- Assignees: Sap SE
- Applicants: Sap SE
- Inventors: Jesper Lind; Suchitra Sundararaman
- Web page: https://patentgrep.com/patent/US11308492B2

## CPC

- G06Q 20/40: Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06F 16/583: using metadata automatically derived from the content
- G06F 18/24: Classification techniques
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/30: Semantic analysis
- G06N 3/02: Neural networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- G06Q 20/045: {using payment protocols involving tickets}
- G06Q 20/047: {using payment protocols involving electronic receipts}
- G06Q 20/389: {Keeping log of transactions for guaranteeing non-repudiation of a transaction}
- G06Q 20/4016: {involving fraud or risk level assessment in transaction processing}
- G06Q 40/12: Accounting
- G06T 7/0002: {Inspection of images, e.g. flaw detection}
- G06T 7/74: {involving reference images or patches}
- G06V 20/95: {Pattern authentication; Markers therefor; Forgery detection}
- G06V 30/18086: {by performing operations within image blocks or by using histograms}
- G06V 30/224: of printed characters having additional code marks or containing code marks
- G06V 30/413: Classification of content, e.g. text, photographs or tables
- G06V 30/414: Extracting the geometrical structure, e.g. layout tree; Block segmentation, e.g. bounding boxes for graphics or text
- G06V 30/416: Extracting the logical structure, e.g. chapters, sections or page numbers; Identifying elements of the document, e.g. authors
- G06V 30/418: Document matching, e.g. of document images
- G06F 16/24564: {Applying rules; Deductive queries}
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 5/041: {Abduction}
- G06T 2207/20061: Hough transform
- G06T 2207/30176: Document
- G06V 30/40: Document-oriented image-based pattern recognition
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

The present disclosure involves systems, software, and computer implemented methods for transaction auditing. One example method includes determining valid pixel-based pattern(s) that are included in valid reference images. Fraudulent pixel-based pattern(s) that are included in fraudulent reference images are determined. A request to classify an image is received. A determination is made as to whether pixel values in the image match a valid pixel-based pattern or a fraudulent pixel-based pattern. In response to determining that the pixel values match a valid pixel-based pattern, a likelihood of classifying the first image as a valid image is increased. In response to determining that the pixel values match a fraudulent pixel-based pattern, a likelihood that the image as a fraudulent image is increased. The image is classified in response to the request as either a valid image or a fraudulent image based on the likelihoods.

## Claims

### Claim 1

A computer-implemented method for auditing financial documents, comprising:
identifying a first corpus of valid reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
identifying a second corpus of fraudulent reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
analyzing first pixel values of first pixels in the valid reference images in the first corpus to determine that at least one valid pixel-based pattern is included in pixels of at least a first threshold percentage of the valid reference images, wherein the valid pixel-based pattern comprises a first edge difference metric, wherein an edge difference metric for any given image measures the degree of variation in the pixel intensities in an edge area, an edge difference metric being generated by:
determining a first mean pixel value for a left edge area of the given image;
determining a second mean pixel value for a right edge area of the given image;
determining a third mean pixel value for an upper edge area of the given image;
determining a fourth mean pixel value for a lower edge area of the given image;
determining a first edge pixel value difference between the third mean pixel value and the first mean pixel value;
determining a second edge pixel value difference between the fourth mean pixel value and the second mean pixel value; and
determining the edge difference metric for the given image by determining the difference between the first edge pixel value difference and the second edge pixel value difference;
analyzing second pixel values of second pixels in the fraudulent reference images in the second corpus to determine that at least one fraudulent pixel-based pattern is included in pixels of at least a second threshold percentage of the fraudulent reference images, wherein the fraudulent pixel-based pattern comprises a second edge difference metric;
receiving a request to classify a first image;
analyzing third pixel values of third pixels included in the first image to determine at least one pixel-based pattern in the third pixels, wherein the pixel-based pattern comprises a third edge difference metric;
determining whether the third edge difference metric matches either a first edge difference metric or a second edge difference metric;
in response to determining that the third edge difference metric matches the first edge difference metric, increasing a first likelihood of classifying the first image as a valid image;
in response to determining that the third edge difference metric matches the second edge difference metric, increasing a second likelihood of classifying the first image as a fraudulent image; and
classifying the first image in response to the request as either a valid image or a fraudulent image based on the first likelihood and the second likelihood.

### Claim 2 (depends on 1)

The method of claim 1, wherein the valid reference images comprise images of printed documents.

### Claim 3 (depends on 2)

The method of claim 2, wherein the valid reference images comprise images of valid receipts.

### Claim 4 (depends on 1)

The method of claim 1, wherein the fraudulent reference images comprise images of documents that are programmatically generated.

### Claim 5 (depends on 4)

The method of claim 4, wherein the fraudulent reference images comprise fraudulent receipt images.

### Claim 6

A system comprising:
one or more computers; and
a computer-readable medium coupled to the one or more computers having instructions stored thereon which, when executed by the one or more computers, cause the one or more computers to perform operations comprising:
identifying a first corpus of valid reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
identifying a second corpus of fraudulent reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
analyzing first pixel values of first pixels in the valid reference images in the first corpus to determine that at least one valid pixel-based pattern is included in pixels of at least a first threshold percentage of the valid reference images, wherein the valid pixel-based pattern comprises a first edge difference metric, wherein an edge difference metric for any given image measures the degree of variation in the pixel intensities in an edge area, an edge difference metric being generated by:
determining a first mean pixel value for a left edge area of the given image;
determining a second mean pixel value for a right edge area of the given image;
determining a third mean pixel value for an upper edge area of the given image;
determining a fourth mean pixel value for a lower edge area of the given image;
determining a first edge pixel value difference between the third mean pixel value and the first mean pixel value;
determining a second edge pixel value difference between the fourth mean pixel value and the second mean pixel value; and
determining the edge difference metric for the given image by determining the difference between the first edge pixel value difference and the second edge pixel value difference;
analyzing second pixel values of second pixels in the fraudulent reference images in the second corpus to determine that at least one fraudulent pixel-based pattern is included in pixels of at least a second threshold percentage of the fraudulent reference images, wherein the fraudulent pixel-based pattern comprises a second edge difference metric;
receiving a request to classify a first image;
analyzing third pixel values of third pixels included in the first image to determine at least one pixel-based pattern in the third pixels, wherein the pixel-based pattern comprises a third edge difference metric;
determining whether the third edge difference metric matches either a first edge difference metric or a second edge difference metric;
in response to determining that the third edge difference metric matches the first edge difference metric, increasing a first likelihood of classifying the first image as a valid image;
in response to determining that the third edge difference metric matches the second edge difference metric, increasing a second likelihood of classifying the first image as a fraudulent image; and
classifying the first image in response to the request as either a valid image or a fraudulent image based on the first likelihood and the second likelihood.

### Claim 7

A computer program product encoded on a non-transitory storage medium, the product comprising non-transitory, computer readable instructions for causing one or more processors to perform operations comprising:
identifying a first corpus of valid reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
identifying a second corpus of fraudulent reference images, wherein each of the images comprises an image of an authentic financial document having a left edge, a right edge, an upper edge, and a lower edge, wherein each of these edges comprise an edge area;
analyzing first pixel values of first pixels in the valid reference images in the first corpus to determine that at least one valid pixel-based pattern is included in pixels of at least a first threshold percentage of the valid reference images, wherein the valid pixel-based pattern comprises a first edge difference metric, wherein an edge difference metric for any given image measures the degree of variation in the pixel intensities in an edge area, an edge difference metric being generated by:
determining a first mean pixel value for a left edge area of the given image;
determining a second mean pixel value for a right edge area of the given image;
determining a third mean pixel value for an upper edge area of the given image;
determining a fourth mean pixel value for a lower edge area of the given image;
determining a first edge pixel value difference between the third mean pixel value and the first mean pixel value;
determining a second edge pixel value difference between the fourth mean pixel value and the second mean pixel value; and
determining the edge difference metric for the given image by determining the difference between the first edge pixel value difference and the second edge pixel value difference;
analyzing second pixel values of second pixels in the fraudulent reference images in the second corpus to determine that at least one fraudulent pixel-based pattern is included in pixels of at least a second threshold percentage of the fraudulent reference images, wherein the fraudulent pixel-based pattern comprises a second edge difference metric;
receiving a request to classify a first image;
analyzing third pixel values of third pixels included in the first image to determine at least one pixel-based pattern in the third pixels, wherein the pixel-based pattern comprises a third edge difference metric;
determining whether the third edge difference metric matches either a first edge difference metric or a second edge difference metric;
in response to determining that the third edge difference metric matches the first edge difference metric, increasing a first likelihood of classifying the first image as a valid image;
in response to determining that the third edge difference metric matches the second edge difference metric, increasing a second likelihood of classifying the first image as a fraudulent image; and
classifying the first image in response to the request as either a valid image or a fraudulent image based on the first likelihood and the second likelihood.