Digital Image Comparison for Identity Document Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for a simple and effective method for individuals to verify the authenticity of identity documents, such as identity cards and passports, which is not as complex as official verification processes, and must be easily implementable by the document holder.
Innovation Solution
A method involving the comparison of digital test images with digital reference images, where areas of interest are identified and subsets are selected and compared, with optional geometric transformations and fingerprint calculations, using a communicating object like a mobile phone with an optical sensor, to determine if the images come from the same source without transmitting the full image file.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the full digital image file is transmitted between the communicating object and the server, then the authentication can be performed, but the security is compromised because the reference image or source can be reconstructed if the communication is intercepted
Solution Approach 1:
The patent extracts only the necessary authentication information (fingerprint of a subset of pixels) from the full image and transmits only this extracted data between the communicating object and server. This allows authentication to proceed without transmitting the complete reference image, thereby preventing reconstruction of the source if communication is intercepted.
Solution Approach 2:
The patent introduces an intermediary cryptographic hash function (fingerprint calculation) that transforms the full image into a condensed representation. This intermediary form enables authentication while preventing direct reconstruction of the original image from transmitted data.
2Ease of operation
If the authentication sequence is predictable, then the authentication process is simpler, but the security is weakened because a third party can intercept and replicate the authentication
Solution Approach 1:
The patent makes the authentication process dynamic by randomly selecting different subsets of pixels and applying different geometric transformations (rotation, scaling, cropping) for each authentication attempt. This ensures that even if a third party intercepts one authentication sequence, they cannot replicate it because the next sequence will be different.
Solution Approach 2:
The patent implements periodic re-authentication with varying parameters, where each authentication cycle uses a different randomly selected subset and transformation. This periodic variation with changing parameters prevents replay attacks while maintaining a relatively simple authentication flow.
3Measurement precision
If a large subset of pixels is selected for comparison, then the authentication accuracy is higher, but the communication bandwidth and processing time increase
Solution Approach 1:
The patent applies partial action by selecting only a subset of pixels from the full image for authentication comparison. This subset is sufficient to determine authenticity while significantly reducing the amount of data that needs to be processed and transmitted, thereby balancing accuracy with efficiency.
4Adaptability or versatility
If geometric transformations (rotation, scaling, cropping) are applied to the subset, then the authentication becomes more robust against different orientations, but the complexity of the comparison process increases
Solution Approach 1:
The patent changes geometric parameters (rotation angle, scale factor, crop position) of the selected pixel subset before comparison. By transforming the subset into a standardized form, the system achieves orientation independence while keeping the comparison process manageable through consistent transformation application.
Data Source
Figure 1
Figure 2~4
AI summary
The invention relates to a method of comparing a test image (IMGT) with a reference image (IMGR), comprising: identifying (100) at least two zones of interest (IZ1, IZ2, IZ3, IZ4) in the reference image. It is essentially characterized in that it consists in: selecting (110) in a random manner one at least of the zones of interest identified in the reference image, selecting (120) a subset (SAR) of the reference image, said subset comprising at least one pixel of said selected zone of interest, selecting (130) at least one subset (SAT) of the test image, said subset of the test image corresponding to said subset of a reference image, and comparing (140) said subset of the test image (SAT) with the corresponding subset (SAR) of the reference image, the method comprising optionally furthermore a step consisting in authenticating (141) said test image as a function of the result of the comparison.