Document Edge Pattern Reconstruction for Secure Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for authenticating secure documents with anti-counterfeit patterns fail to accurately detect deformed or randomly positioned patterns, leading to low authentication success rates and high false alerts.
Innovation Solution
A method involving the generation of a combined digital image by merging edge areas of a document to form a complete anti-counterfeit pattern, using a deep learning model like YOLO, and verifying additional information to enhance authentication accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the current solution checks anti-counterfeit patterns at edge(s) of digital copy, then the authentication process is simple, but the authentication success rate is low due to document fatigue and deformation
Solution Approach 1:
The authentication process is segmented into multiple stages: initial edge pattern detection, followed by comprehensive full-image pattern detection. This segmentation allows the system to handle complex authentication by breaking it down into manageable steps, improving success rate without overwhelming complexity at any single stage
Solution Approach 2:
The system performs preliminary detection at the edge areas first to quickly identify potential matches, then uses this information to guide more comprehensive full-image analysis. This preliminary action reduces the search space and improves overall authentication reliability
2Reliability
If anti-counterfeit patterns are randomly positioned during manufacturing, then each document is unique and secure, but detection becomes more difficult due to varying positions and deformations
Solution Approach 1:
The system transitions from one-dimensional edge-only detection to two-dimensional full-image analysis. By examining the entire image space rather than just edges, the system can locate randomly positioned patterns anywhere in the document, accommodating random positioning while maintaining detection capability
Solution Approach 2:
The detection parameters are adjusted to account for document fatigue and deformation. The system uses flexible matching criteria that can tolerate position variations and shape deformations, allowing accurate detection of randomly positioned patterns even when distorted
3Shape
If only partial anti-counterfeit patterns are visible at edge lines, then the document structure is maintained, but the patterns are incomplete and illegible for authentication
Solution Approach 1:
The system merges information from multiple sources: partial patterns detected at edge areas are combined with patterns detected in the full image. This merging creates a complete view of the anti-counterfeit patterns, achieving both document structure preservation and pattern completeness for accurate authentication
Data Source
AI summary
The present invention relates to a method for authenticating a document comprising at least one line of anti-counterfeit patterns spaced from each other, the positions of the anti-counterfeit patterns being random, and a partial anti-counterfeit pattern being provided at an edge line of said document, the method comprising the following steps of:receiving a digital image of said document, said digital image comprising said partial anti-counterfeit pattern in a first edge area corresponding to said edge line;selecting said first edge area;copy-pasting said first edge area adjacent to a second edge area of the digital image to generate a combined digital image, said second edge area being opposite to said first edge area, wherein content, which is comprised in said second edge area and located on the same line as said partial anti-counterfeit pattern, and said partial anti-counterfeit pattern jointly form a combined pattern;authenticating said document by taking into account said combined pattern.