Scanned Image Misinterpretation Correction for Accurate ACR Data

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Automated character recognition (ACR) algorithms introduce typographical errors in scanned images due to visual misinterpretations, leading to discrepancies in interaction data that can erroneously flag interactions as unauthorized, requiring manual intervention.

Innovation Solution

A machine-learning model is employed to detect and resolve visual misinterpretations in scanned images by applying adjustments to the interaction data or the image itself, improving the accuracy of ACR algorithms and enabling automated correction of typographical errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If automated character recognition algorithms are used to extract interaction data from scanned images, then processing efficiency is improved, but measurement precision deteriorates due to visual misinterpretations and typographical errors

Engineering Contradiction:

Improveprocessing efficiencyVSAvoiddata extraction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system compares extracted interaction data against expected values and uses machine learning models to analyze discrepancies. When visual misinterpretations are detected, the system automatically generates corrections by analyzing the scanned image and applying context-aware adjustments to the extracted data, creating a feedback loop that continuously refines accuracy without manual intervention

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

A machine learning model serves as an intermediary between the automated character recognition algorithm and the final interaction data. This intermediary layer analyzes discrepancies, determines whether they result from visual misinterpretations, and applies appropriate corrections before the data is used for interaction processing

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If manual review of interaction data is performed to verify accuracy, then measurement precision is improved, but productivity deteriorates due to inefficient manual processing

Engineering Contradiction:

Improvedata verification accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system performs self-verification by automatically comparing extracted data against expected values and using machine learning models to detect and correct visual misinterpretations. The system identifies its own errors and corrects them autonomously, eliminating the need for manual review while maintaining high accuracy

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system continuously monitors extracted data for discrepancies and automatically initiates correction processes when visual misinterpretations are detected. This self-correcting feedback mechanism ensures high precision without requiring manual verification, thereby maintaining both accuracy and processing speed

Inventive Principle:

Principle #23Feedback

3Productivity

If automated processing is implemented to handle interactions, then productivity is improved, but reliability deteriorates due to false positive unauthorized interaction flags

Engineering Contradiction:

Improveinteraction processing volumeVSAvoidinteraction validation accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system compares extracted interaction data against expected values and uses machine learning models to analyze discrepancies. When visual misinterpretations cause false flags, the feedback mechanism detects these errors and automatically corrects the interaction data, preventing legitimate interactions from being blocked while maintaining security

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The machine learning model acts as an intermediary that filters between automated extraction and final validation decisions. It analyzes discrepancies in context and determines whether they represent actual security threats or merely visual misinterpretations, thereby reducing false positives while maintaining processing throughput

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20260038290A1Automatically detecting and resolving visual misinterpretations of scanned images by a computer

Publication Date: 2026.02.05 TRUIST BANK

AI summary

In some examples, a system can use machine learning to automatically detect and resolve a visual misinterpretation of a scanned image generated by an automated character recognition (ACR) algorithm. For example, the system can execute a machine-learning model on interaction data extracted from an image of a physical document for initiating an interaction between entities. The interaction data can include a discrepancy such that the interaction data is different from one or more expected values. The machine-learning model can determine whether the discrepancy was caused by a visual misinterpretation of the image when the ACR algorithm was applied to the image to extract the interaction data. In response to determining that the discrepancy was caused by the visual misinterpretation, the machine-learning model can apply an adjustment to the interaction data to resolve the visual misinterpretation. Applying the adjustment can generate updated interaction data usable to initiate the interaction.