Real-Time Transaction Fraud Detection via Image Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current systems lack efficient methods for real-time fraud detection in transactions, often relying on manual identification and delayed processing, which can lead to erroneous claims and inefficiencies.

Innovation Solution

A method and system that utilize image processing and real-time transaction data analysis, where a computing system receives and analyzes images captured during transactions to detect features, generating a designation of 'verified' or 'unverified' to determine fraudulent activity, and stores this information for instant fraud claim initiation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If manual identification and delayed processing are used for fraud detection, then system complexity is reduced, but detection speed and accuracy deteriorate

Engineering Contradiction:

Improvesystem complexityVSAvoiddetection speed

Core Design Contradiction:

Device complexityVSSpeed

Solution Approach 1:

The patent replaces manual fraud identification with automated image processing and machine learning algorithms. The system uses computer vision to analyze transaction images, extract features, and automatically determine fraud likelihood, eliminating the need for manual review while significantly improving detection speed and consistency.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system performs preliminary fraud analysis by processing images and transaction data in real-time during the transaction flow. By conducting fraud detection before transaction completion, the system enables immediate intervention for suspicious transactions, improving both speed and accuracy of fraud identification.

Inventive Principle:

Principle #10Preliminary action

2Loss of energy

If manual fraud detection is used, then processing costs are reduced, but detection accuracy and fraud prevention capability worsen

Engineering Contradiction:

Improveprocessing costsVSAvoiddetection accuracy

Core Design Contradiction:

Loss of energyVSReliability

Solution Approach 1:

The system implements self-service fraud detection by automatically analyzing transaction images and data without requiring human intervention. The machine learning models independently evaluate transactions, generate fraud scores, and trigger appropriate responses, reducing labor costs while improving detection accuracy through consistent automated analysis.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent introduces an intermediary image processing layer that automatically extracts features from transaction images and translates them into structured data for fraud analysis. This intermediary system bridges the gap between raw transaction data and fraud detection algorithms, improving accuracy while maintaining cost efficiency through automated processing.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Measurement precision

If real-time image analysis is implemented, then fraud detection accuracy improves, but computational requirements and system complexity increase

Engineering Contradiction:

Improvefraud detection accuracyVSAvoidcomputational requirements

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the fraud detection process into distinct modular components: image preprocessing, feature extraction, transaction data processing, fraud scoring, and decision-making. Each module handles specific computational tasks independently, improving accuracy through specialized processing while managing system complexity through modular architecture that allows selective deployment.

Inventive Principle:

Principle #1Segmentation

4Reliability

If comprehensive transaction data collection is performed, then fraud detection reliability improves, but data processing time and system complexity increase

Engineering Contradiction:

Improvefraud detection reliabilityVSAvoiddata processing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary organization and pre-processing of transaction data and images as they are generated, structuring them for immediate analysis. By preparing data in advance with proper formatting and feature extraction, the system enables rapid fraud detection without requiring extensive post-capture processing, thus improving both reliability and reducing processing time.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12190326B2Systems and methods for digitized proof of transactions

Publication Date: 2025.01.07 WELLS FARGO BANK NA

AI summary

A method can include receiving first data corresponding to a transaction between a merchant and a customer; receiving second data captured at substantially the same time as the transaction and at a location corresponding to the transaction; generating a designation for the transaction based on a first match between the good identified in the first image and the identifying information and a second match between the depiction of the customer and a profile image of the customer; storing third data comprising at least one of the plurality of images and a unique identifier assigned to the transaction; and linking the second data to the third data; populating a claim with at least some of the second data and the third data according to a selection of a user interface element.