Dual-Image Face Recognition for Infrared Photo Spoof Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition technologies are vulnerable to being cracked by using infrared photos with registered face images, as they cannot distinguish between real faces and infrared photos.
Innovation Solution
A face recognition device and method that captures two images of a recognition target, analyzes the first image to determine if it is a real face based on image conditions, and then performs facial feature analysis on the second image to match with registered data, using different illumination levels or types to differentiate between real faces and infrared photos.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single infrared image is used for face recognition, then the recognition process is simple and fast, but the system cannot distinguish between real faces and infrared photos, making it vulnerable to cracking
Solution Approach 1:
The patent divides the face recognition process into two distinct stages: first capturing a color image without infrared illumination to verify it's a real face, then capturing an infrared image for actual recognition. This segmentation allows the system to separately handle verification and recognition functions, preventing infrared photo attacks while maintaining recognition efficiency.
Solution Approach 2:
The system performs preliminary verification by capturing a color image before proceeding to infrared recognition. This preliminary action checks whether the target is a real face by analyzing color information and facial features, preventing fraudulent infrared photos from entering the recognition pipeline.
2Reliability
If two images (color and infrared) are captured and analyzed, then the anti-cracking capability is improved, but the device complexity and processing time increase
Solution Approach 1:
The camera module is designed to perform multiple functions: capturing both color images and infrared images using the same hardware platform. The infrared camera can switch between different imaging modes, and the processor handles both color and infrared image processing, reducing the need for separate dedicated hardware for each function.
Solution Approach 2:
The patent combines the color camera and infrared camera into a single integrated camera module that shares common components such as the processor, memory, and control circuitry. This merging approach reduces overall system complexity while maintaining the dual-image capture capability needed for anti-cracking protection.
3Reliability
If two images are captured and analyzed, then the anti-cracking capability is improved, but the processing time and energy consumption increase
Solution Approach 1:
The color image capture and analysis is performed as a preliminary step before infrared recognition. By quickly verifying the target is a real face through color image analysis, the system can early-reject infrared photos and proceed directly to infrared recognition only when necessary, reducing overall processing time for fraudulent attempts.
Solution Approach 2:
The system uses periodic or sequential action by first capturing a color image, then conditionally capturing an infrared image based on the color image analysis results. This periodic approach ensures that full dual-image processing is only performed when needed, reducing average processing time while maintaining security.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively prevents unauthorized access by ensuring that only real faces are recognized, thereby enhancing security and preventing fraudulent use of infrared photos.
Implementation Method 1
The infrared light emitting module is coupled to the processor. The camera module includes an infrared camera. The infrared camera is configured to non-simultaneously obtain the first image and the second image. When the infrared camera obtains the first image, the infrared light emitting module simultaneously emits a first illumination light to the recognition target.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A face recognition device (100) and a face recognition method are provided. The face recognition device (100) includes a camera module (120), a processor (110) and a memory (130). The camera module (120) obtains a first image of a recognition target (200) and obtains a second image of the recognition target (200). The processor (110) analyzes whether the first image meets an image condition to determine whether the recognition target (200) is a real face. The memory (130) stores a face database. When the processor (110) determines that the first image meets the image condition, the processor (110) analyzes a plurality of facial features in the second image, and the processor (110) compares whether the second image matches face registration data in the face database according to the facial features to determine whether the second image passes recognition.