Validation Engine for OCR Conflict Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optical character recognition (OCR) systems face challenges in accurately translating characters from images, especially in poor weather conditions or low light, leading to inaccurate identification and billing issues in vehicle charging environments, despite high confidence ratings.
Innovation Solution
A validation engine is introduced to analyze conflicting image identifications from OCR systems, determining actions based on confidence levels and predetermined operating parameters, which may involve manual verification or notification to a billing system, to improve accuracy and resolve conflicts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If OCR systems are used to translate characters from images in challenging conditions, then identification speed is improved, but identification accuracy deteriorates
Solution Approach 1:
The patent combines multiple OCR engine outputs and integrates them with confidence value analysis and conflict resolution logic into a unified validation system. This merging of multiple identification sources and validation mechanisms improves overall identification accuracy while maintaining processing efficiency through automated conflict resolution.
Solution Approach 2:
The system implements feedback through confidence value analysis where OCR engine outputs are evaluated against predetermined confidence thresholds. When confidence values indicate potential errors or conflicts arise between multiple OCR engines, the system triggers validation actions including manual verification requests, creating a feedback loop that continuously improves identification accuracy.
2Reliability
If multiple OCR engines are deployed to improve identification reliability, then system complexity increases
Solution Approach 1:
The validation engine serves multiple functions: it receives and processes outputs from multiple OCR engines, analyzes confidence values, resolves conflicts between different identifications, and determines appropriate validation actions. This multi-functional design consolidates what would otherwise require separate systems into a single unified component, managing complexity while improving reliability.
Solution Approach 2:
The system implements self-service through automated conflict resolution mechanisms. When conflicts are detected between multiple OCR engine outputs, the validation engine automatically analyzes confidence values and determines whether manual verification is needed, reducing the need for constant human intervention and simplifying system operation despite the presence of multiple OCR engines.
3Measurement precision
If manual verification is performed for conflicting identifications, then identification accuracy is improved, but processing time increases
Solution Approach 1:
The system applies partial verification by requesting manual validation only when confidence values fall below predetermined thresholds or when conflicts arise between multiple OCR engine outputs. For high-confidence, non-conflicting identifications, the system proceeds directly to billing without manual review, thus improving accuracy where needed while minimizing processing time delays.
Solution Approach 2:
The validation engine dynamically adjusts the verification process based on confidence value parameters. By analyzing confidence levels and conflict severity, the system determines the appropriate level of validation required, changing the verification parameter from full manual review to automated processing based on the specific case, thereby optimizing the balance between accuracy and processing time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A validation engine for determining an action to be performed in response to conflicting image identifications being generated by an optical character recognition system, the validation engine comprising:an analysis component for receiving the conflicting image identifications from the optical recognition system, each image identification having a generated confidence value; the analysis component analysing the generated confidence values to determine an identification conflict between each of the received image ident ificat ions; and the analysis component analysing the generated conflict to identify whether the identification conflict falls within a predetermined range of operating parameters and, in dependence of where in the range the identification conflict falls determining an action to be performed.