Physical ID Document Fraud Labeling Through Image Alignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fraud detection systems for physical identification documents face challenges in scalability, accuracy, and efficiency due to the need for expert knowledge and manual labeling, leading to inconsistent and biased results, which are costly and time-consuming.

Innovation Solution

A computerized system and method that employs a machine learning classification model to generate and deploy fraud detection training data by aligning and cropping images, associating fraud detection questions, and receiving responses to improve labeling accuracy and speed, enabling junior analysts to achieve expert-level performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual labeling by expert analysts is used to ensure high-quality training data, then labeling accuracy is improved, but productivity is reduced and costs increase

Engineering Contradiction:
Improvelabeling accuracyVSAvoidlabeling speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The system performs preliminary actions by automatically generating synthetic fraudulent documents and pre-labeling training data before human analysts need to review it. This prepares high-quality training data in advance, reducing the manual work required while maintaining accuracy standards.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system creates synthetic copies of authentic documents with fraudulent modifications. These synthetic copies serve as training examples that mimic real fraud cases without requiring actual fraudulent documents, enabling scalable training data generation while preserving labeling quality.

Inventive Principle:
Principle #26Copying

2Measurement precision

If more training data is collected to improve model accuracy, then measurement precision is improved, but loss of time increases due to manual review requirements

Engineering Contradiction:
Improvefraud detection accuracyVSAvoidtime for manual labeling
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system generates synthetic copies of documents with various fraud modifications, creating large volumes of training data automatically. This eliminates the time-consuming manual collection and labeling process while providing diverse training examples for improved model accuracy.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system performs self-service by automatically generating and labeling training data without requiring human analysts for each sample. The automated system handles data generation, labeling, and quality control, freeing human analysts to focus only on oversight and validation tasks.

Inventive Principle:
Principle #25Self-service

3Reliability

If human analysts manually review and label documents to reduce bias, then reliability is improved, but productivity decreases

Engineering Contradiction:
Improveconsistency of labelingVSAvoidthroughput of labeling
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system performs self-service by automatically generating consistently labeled training data through algorithmic processes. This eliminates human variability and bias while maintaining high throughput, as the automated system can process large volumes of data without fatigue or inconsistency.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system changes the approach from human-based labeling to algorithmic labeling, fundamentally altering the parameter of who performs the labeling. This transition maintains reliability through consistent algorithmic application while dramatically increasing productivity through automation.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4592970A1Methods and systems for generating and deploying fraud detection training data for physical identification documents
Publication Date: 2025.07.30 ONFIDO LTD
  • EP4592970A1 patent drawingFigure 1
  • EP4592970A1 patent drawingFigure 2
  • EP4592970A1 patent drawingFigure 3

AI summary

Described herein are computerized methods and systems for generating and deploying fraud detection training data for physical identification documents. A server identifies a review image depicting a physical identification document to be validated, the document comprising areas of interest each associated with a fraud signal. The server generates a dataset comprising reference images, each depicting a reference document. The server aligns document features depicted in the review images and the reference images and crops each aligned image. The server identifies a fraud detection question for the cropped review image and displays the review image, the reference images, and the question in a user interface. The server receives a response to the fraud detection question and determines accuracy of the question based upon the response. The server labels the review image as genuine or fraudulent when the accuracy of the question is above a threshold.