Coin Recognition Templates Using Virtual Boundary Lines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing coin recognition systems struggle to accurately identify soiled coins, particularly bimetal coins, immediately after a new series is introduced, as they require physical templates from actual coins, which are often unavailable or insufficiently soiled.
Innovation Solution
A feature information generation device and method that generates virtual boundary lines on coin images to create synthetic images, reducing the influence of boundary line soiling, allowing for accurate recognition even when physical soiled coins are unavailable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If templates are created from actual soiled coins, then recognition accuracy for soiled coins improves, but availability of templates is insufficient immediately after new coin introduction
Solution Approach 1:
The patent creates synthetic coin images that copy and simulate the appearance of soiled coins through image processing techniques. Virtual soiling is applied to clean coin images to generate templates that resemble actual soiled coins, enabling template creation without requiring physical soiled coins to be available.
Solution Approach 2:
The system performs preliminary creation of recognition templates by synthesizing soiled coin images from clean coin images before actual soiled coins become available in circulation. This allows the recognition system to be prepared in advance with appropriate templates for new coin series.
2Ease of manufacture
If templates are created from new unused coins, then template creation is feasible immediately after introduction, but recognition accuracy for soiled coins deteriorates
Solution Approach 1:
The patent uses image processing to copy the appearance characteristics of soiled coins and apply them to clean coin images. This creates synthetic templates that retain the structural information from new coins while incorporating visual characteristics of soiling, enabling both easy template creation and accurate soiled coin recognition.
Solution Approach 2:
The system changes visual parameters of clean coin images by applying virtual soiling effects, transforming them into synthetic soiled coin images. This allows templates to be created from easily available new coins while achieving the recognition accuracy needed for soiled coins through parameter modification.
3Reliability
If boundary line soiling is included in templates, then authenticity determination is more comprehensive, but false positives increase due to pattern distortion
Solution Approach 1:
The patent applies different processing qualities to different regions of the coin image. Boundary line regions with soiling are processed differently from the main body of the coin, allowing the system to account for soiling effects locally without allowing them to distort the overall feature extraction and authenticity determination.
Data Source
Figure 1~3
Figure 4~5
Figure 6~7
AI summary
The feature information generation device of the present disclosure generates feature information for coin recognition from an image of a coin. The feature information generation device includes: an acquisition unit configured to acquire an actual coin image which is an image of an actual coin, the actual coin including a first surface part formed from a first material and a second surface part formed from a second material different from the first material; and a control unit configured to generate a synthetic coin image to which a virtual boundary line is provided along a boundary between the first surface part and the second surface part in the actual coin image and generate the feature information based on the synthetic coin image.