Identity Document Forgery Detection Using Neural Network Image Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine-based identity verification methods cannot detect tampering of visual security elements in identity documents, allowing fraudsters to reuse genuine documents with swapped security elements.
Innovation Solution
A method using a processor in a security device to select regions of interest, create a composite image, apply digital image filtering to reveal geometrical objects, and utilize a trained neural network to detect forgery based on these objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine-based verification using OCR and digital processing is used, then verification speed and automation are improved, but the ability to detect tampering of visual security elements deteriorates
Solution Approach 1:
The method segments the identity document image into multiple regions of interest, specifically focusing on the visual security element and its surrounding areas. This segmentation allows the system to apply specialized processing to detect tampering patterns that would be missed in full-document processing, thereby maintaining high automation while improving detection reliability.
Solution Approach 2:
The system applies digital image filtering that enhances color gradient and laplacian measures to detect changes in color patterns and shading within the visual security element. These color analysis techniques reveal tampering artifacts that automated OCR-based systems cannot detect, thus improving reliability while preserving automation.
2Measurement precision
If digital image filtering and neural networks are applied to detect geometrical objects, then detection precision for forgery is improved, but device complexity increases
Solution Approach 1:
The system performs preliminary actions by pre-defining regions of interest and pre-processing the image data through digital filtering before feeding it to the neural network. This preliminary processing extracts and enhances the most relevant features, allowing the neural network to focus only on classification rather than feature extraction, thereby improving precision while managing computational complexity.
Solution Approach 2:
The digital image filtering acts as an intermediary between the raw image input and the neural network classifier. This intermediary layer processes the image to reveal geometrical objects and tampering patterns, transforming the complex task of forgery detection into a more manageable classification problem, thus improving precision without proportionally increasing overall system complexity.
Data Source
AI summary
Provided is a method for detecting a forgery of an identity document including a visual security element. A neural network is trained with a training data set such that an input to the dedicated neural network is a filtered image for a given identity document and an output of the dedicated neural network is an indicator of the forgery or not of said given identity document, wherein said output is based on geometrical objects in said input filtered image created by a replacement or a displacement of a security element in said given identity document and revealed by said digital image filtering, thereby producing a set of parameters for the dedicated neural network with which the processor of the security device has been programmed. Other embodiments disclosed.