Serial Number Recognition Parameter Tuning for Paper Sheet Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The development of serial number recognition programs for paper sheets, such as paper currencies, requires significant man-hours due to the need for individual customization for each recognition target, particularly across different countries.
Innovation Solution
A parameter determination apparatus and program utilize a Genetic Algorithm (GA) to optimize parameters like binarization thresholds, dilation amounts, and distance thresholds, reducing the manual effort needed to develop the serial number recognition program by automating the parameter setting process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a serial number recognition program is individually generated for each recognition target to improve recognition performance, then recognition performance is improved, but the development time and man-hours increase
Solution Approach 1:
The patent creates a universal serial number recognition program that can handle multiple recognition targets (different countries, currencies, paper sheets) through a single framework. The program uses a database of serial number images and automated parameter determination to adapt to different targets without requiring individual program generation for each, thus maintaining high recognition performance while reducing development time.
Solution Approach 2:
The patent employs automated parameter determination that adjusts recognition parameters based on the specific characteristics of each recognition target. Instead of creating separate programs, the system modifies parameters such as image processing thresholds, region of interest coordinates, and recognition criteria dynamically according to the target type, enabling one program to optimize performance across multiple applications.
2Measurement precision
If manual parameter setting is used to optimize recognition for each target, then recognition accuracy improves, but development complexity and man-hours increase
Solution Approach 1:
The patent implements a self-service parameter determination system that automatically optimizes recognition parameters without requiring manual intervention. The system uses automated image processing to analyze serial number characteristics and independently determines optimal parameters for each recognition target, eliminating the need for manual parameter setting while maintaining high recognition accuracy and reducing development complexity.
Solution Approach 2:
The patent performs preliminary analysis of serial number image characteristics during the development phase to establish comprehensive parameter settings in advance. By pre-determining parameters based on analyzed characteristics of different serial number formats and layouts, the system prepares optimized recognition configurations before actual recognition operations, avoiding the need for manual parameter adjustment during deployment.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A parameter determination apparatus 30 includes a generation unit (31), an identification unit (32), and an evaluation index calculation unit (33). The generation unit (31) generates parameter values of a program that is used when identifying character present regions, and the identification unit (32) identifies the character present regions by simulation by using the parameter values generated by the generation unit (31). Further, the evaluation index calculation unit (33) calculates fitness ftotal of the parameter values based on the character present regions that are identified by the identification unit (32).