Document Identifier Validation for Real-Time Transaction Fraud Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online transaction systems struggle to detect fraudulent electronic transactions in real-time, allowing fraudsters to execute unauthorized activities before their accounts are invalidated, and sophisticated fraud involves using valid-looking documents with manipulated identifiers.
Innovation Solution
A multi-modal approach using neural networks for document classification, segmentation, and validation to automatically identify and verify identifiers in uploaded documents, combining machine learning and software engineering for real-time fraud detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional transaction verification methods are used, then system complexity is low, but fraud detection accuracy and real-time capability are insufficient
Solution Approach 1:
The verification system is segmented into multiple specialized neural networks: a document verification network to authenticate document authenticity, an identifier extraction network to locate and extract identifier fields, and an OCR network to recognize character values. This segmentation allows each component to specialize in one aspect of fraud detection, improving overall accuracy while managing complexity through modular architecture.
Solution Approach 2:
The patent introduces an intermediary validation layer that sits between traditional verification systems and the transaction processing. This intermediary uses multiple neural networks to first verify document authenticity, then extract and validate identifiers against institutional databases, providing a bridge between simple verification methods and complex fraud detection requirements.
2Reliability
If real-time fraud detection is implemented, then transaction security is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary document verification and identifier extraction in sequence before final transaction authorization. The document verification network authenticates the document first, then the segmenter extracts identifiers, and only after these preliminary steps are completed does the system query institutional databases for final validation. This staged approach allows real-time security while optimizing processing time by avoiding unnecessary computations on obviously fraudulent documents.
3Measurement precision
If multiple neural networks are used for comprehensive document verification, then detection precision is improved, but device complexity and computational overhead increase
Solution Approach 1:
The verification task is segmented across three specialized neural networks: document verification network for authenticity checking, segmenter network for identifier location and extraction, and OCR network for character recognition. Each network is optimized for its specific function, achieving high detection precision while managing complexity through functional specialization and modular architecture.
Data Source
AI summary
A method receives an electronic image and uses the image as an input to a neural network. Based on a determination that the image represents a document, the method uses the image as an input to another neural network to identify a portion of the document containing an identifier. The method extracts the identifier by performing character recognition on the identified portion and determines whether the identifier is valid by using a validation API to determine whether the identifier is associated with a valid account at an institution. Based on a determination that the identifier is associated with a valid account, the method authorizes a transaction associated with the identifier. Based on a determination that the identifier is not associated with a valid account, the method denies the transaction. The first neural network classifies the electronic image into one of multiple valid document types and an invalid document type.