Spectrum Multi-Band Fusion Face Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition systems struggle to accurately differentiate between real and fake human faces, especially under poor lighting conditions and are affected by high-definition 3D masks and facial feature changes through makeup, leading to security challenges and high costs associated with infrared cameras and spectrometers.
Innovation Solution
A face recognition monitoring system based on spectrum and multi-band fusion using a silicon-based detector divided into units with visible-near-infrared and RGGB image acquisition areas, allowing for the collection of multi-spectral data and images through a single camera, which reduces costs and enhances recognition accuracy by distinguishing real human skin.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If infrared cameras are used to eliminate the effect of poor light conditions, then the face anti-spoofing capability is improved, but the system cost increases significantly
Solution Approach 1:
The patent combines visible light imaging and near-infrared imaging into a single integrated camera system with a unified optical path. The silicon-based detector simultaneously captures both visible and near-infrared spectral information, eliminating the need for separate infrared cameras and reducing overall system cost while maintaining anti-spoofing capability.
Solution Approach 2:
The patent designs a multi-functional camera system where a single device performs both visible light face recognition and near-infrared anti-spoofing detection. The silicon-based detector is configured to respond to both visible and near-infrared wavelengths, allowing one component to serve multiple functions and reduce system complexity.
2Measurement precision
If high-precision spectrometers are used to obtain spectral feature information, then the accuracy of identifying real and fake faces is improved, but the system cost increases and practical application becomes difficult
Solution Approach 1:
The patent replaces expensive high-precision spectrometers with a cost-effective silicon-based detector that can capture spectral information in the visible and near-infrared ranges. While not as precise as high-end spectrometers, this detector provides sufficient spectral discrimination capability for anti-spoofing applications at a much lower cost, making practical deployment feasible.
Solution Approach 2:
The patent shifts the operational parameters from requiring full-spectrum high-precision spectrometry to utilizing specific wavelength ranges (visible and near-infrared) where silicon detectors have optimal performance. By focusing on these specific spectral bands, the system achieves adequate discrimination between real and fake faces using cheaper components.
3Measurement precision
If RGGB cameras are used for face recognition, then the imaging function is improved, but the accuracy of face anti-spoofing under poor light conditions deteriorates
Solution Approach 1:
The patent extends the spectral dimension by incorporating near-infrared sensing capability into the camera system. While RGGB cameras provide good visible light imaging, the addition of near-infrared detection capability allows the system to operate effectively in poor light conditions and perform spectral-based anti-spoofing by utilizing information from an additional spectral dimension.
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
The system effectively identifies real and fake human faces with high accuracy, preventing spoofing attacks and improving recognition security while being cost-effective by integrating visible-near infrared and RGGB image data for dual recognition.
Implementation Method 1
Spectral modulation film layers of different band are respectively formed on the three areas by etching, so that the reflection multi-spectral data of the human face in visible-near-infrared band can be obtained
Implementation Method 2
Spectral modulation film layers of different band are respectively formed on the three areas by etching, so that the reflection multi-spectral data of the human face in visible-near-infrared band can be obtained
Implementation Method 3
a silicon-based detector is divided into multiple units, and each unit has three areas... a face spectral image, a face image in the near-infrared band and a RGGB face image can be obtained through one camera
Data Source
AI summary
A face recognition monitoring system based on spectrum and multi-band fusion, including a spectrum camera, a first module for acquiring a face spectral image, a second module for preprocessing data of the face spectral image, a face spectral image database and a third module for recognizing the face spectral image. The spectrum camera includes an optical lens and a silicon-based detector. The silicon-based detector includes a photoelectric conversion substrate and a filter film arranged thereon. The filter film includes N units each including a visible spectrum sensing area, a near-infrared spectral image sensing area and a RGGB image acquisition area. The N units cover all pixels on the photoelectric conversion substrate. A recognition method using the above system is also provided.