Biometric Face Image Manipulation Detection via Segmented Feature Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The reliability of automated access control systems is compromised by the ease with which digital reference facial images can be manipulated, making it difficult to ensure the authenticity of identity documents, as manipulated images can biometrically fit multiple individuals.
Innovation Solution
A device and method that compares digital reference facial images from identity documents with captured facial images, using biometric features and manipulation indicators to detect image manipulation, such as morphing, through statistical and neural network evaluations, to determine the likelihood of image tampering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If digital reference facial images are stored in identity documents for automated access control, then convenience and speed of identification are improved, but reliability deteriorates due to ease of image manipulation
Solution Approach 1:
The patent applies preliminary action by capturing and storing multiple facial images of the same person at different times during the identity document issuance process. These pre-captured images serve as a reference baseline before any potential manipulation can occur, enabling future detection of unauthorized changes when the stored reference image is compared against newly captured images
2Reliability
If image manipulation detection is implemented to improve reliability, then authenticity verification is improved, but device complexity increases
Solution Approach 1:
The patent uses copying by storing multiple copies of facial images captured at different times in the identity document. Instead of implementing complex real-time manipulation detection algorithms, the system creates simple image copies that can be directly compared, reducing computational complexity while maintaining reliable detection capability
Solution Approach 2:
The patent applies partial action by focusing comparison on specific facial features and key regions rather than analyzing the entire image in detail. This selective approach reduces the computational burden and simplifies the detection system while still achieving reliable manipulation detection
3Measurement precision
If multiple biometric comparisons are performed to detect manipulation, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies segmentation by dividing the facial image into multiple regions or features (such as eyes, nose, mouth, or other distinctive facial characteristics) and performing comparisons on these segmented elements. This allows parallel processing of different facial features, improving detection accuracy through multiple comparison points while reducing overall processing time through concurrent evaluation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a device (100) for detecting image manipulation in a digital reference facial image of a person. The device (100) comprises a document reader (101) configured to read the digital reference facial image of the person from an identity document of the person. The device (100) further comprises an image sensor (103) configured to capture a digital facial image of the person. The device (100) further comprises a detection device (105) configured to capture a biometric reference feature in the digital reference facial image, to compare the biometric reference feature with the biometric feature to obtain a manipulation indicator, and to detect the image manipulation in the digital reference facial image based on the manipulation indicator.The device (100) may also include a database (107) in which a biometric comparison reference feature is stored.