3D Face Authentication Using Depth and Thermal Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current facial recognition methods are plagued by low recognition rates and instability due to ambient light conditions and the difficulty in distinguishing between real and artificial faces, particularly in environments with varying illumination and the presence of shadows or 2D images.
Innovation Solution
A 3D face identity authentication method utilizing depth images and 2D images, which calculates a 3D texture image, projects it onto a 2D plane, extracts feature information, and compares it with a reference face to determine similarity, while also incorporating eye-gaze detection and liveness verification to ensure authenticity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D face recognition is used, then the device complexity is low, but the recognition accuracy and stability are poor under varying light conditions
Solution Approach 1:
The patent transitions from 2D face recognition to 3D face recognition by introducing depth information through structured light projection and time-of-flight measurement. This dimensional change enables accurate face geometry capture that is independent of ambient lighting conditions, resolving the contradiction between recognition accuracy and device complexity by adding only specific depth-sensing components rather than a complete system overhaul
Solution Approach 2:
The patent changes the measurement parameter from 2D color intensity to 3D depth distance. By measuring the time of flight of projected structured light or the phase shift of modulated light, the system obtains depth information that remains stable under varying illumination, thus improving recognition accuracy while maintaining reasonable device complexity through parameter transformation
2Reliability
If thermal infrared imaging is used to detect real faces, then the reliability against artificial features is improved, but the image resolution becomes too low for accurate recognition
Solution Approach 1:
The patent merges two imaging modalities: structured light depth imaging (which provides high-resolution 3D geometry) and thermal infrared imaging (which provides liveness verification). The depth image captures detailed facial geometry for accurate recognition, while the thermal image verifies that the face is real and not an artificial replica, thus resolving the contradiction by combining the strengths of both methods
Solution Approach 2:
The patent creates a composite authentication approach by combining depth information from structured light with thermal information from infrared imaging. This composite method uses depth data for high-resolution feature extraction and thermal data for liveness verification, achieving both high resolution and high reliability simultaneously
3Reliability
If near-infrared imaging is used, then the stability under ambient light is improved, but the system remains vulnerable to artificial feature deception
Solution Approach 1:
The patent makes the authentication system multi-functional by integrating three capabilities: 3D geometry capture via structured light, high-resolution imaging, and liveness detection via thermal sensing. This universal system can simultaneously achieve stable recognition under ambient light and detect artificial features, as the thermal modality specifically targets liveness verification while the depth modality provides geometric authenticity
4Measurement precision
If 3D depth imaging and 2D imaging are combined, then the recognition accuracy and anti-deception capability are improved, but the device complexity increases
Solution Approach 1:
The patent segments the authentication process into distinct functional modules: structured light projection for depth measurement, time-of-flight or phase-shift processing for 3D reconstruction, thermal infrared sensing for liveness verification, and multi-modal fusion for final authentication. This segmentation allows each component to be optimized independently and integrated systematically, managing device complexity while achieving high recognition accuracy and anti-deception capability
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
This approach enhances the accuracy and reliability of facial recognition by reducing false positives and improving recognition rates under varying light conditions, providing a comprehensive solution for tasks like unlocking and payment.
Implementation Method 1
a first camera obtains a current face depth image including the target face
Implementation Method 2
a current face three-dimensional texture image is calculated based on the current face depth image and the current face original image
Data Source
AI summary
The present application provides an identity authentication method and an apparatus. The method may include obtaining a sequence of depth images containing a target face and a sequence of original two-dimensional (2D) images containing the target face, and performing identity authentication. The identity authentication may be conducted by: calculating a target face three-dimensional (3D) texture image according to the depth images containing the target face and the original 2D images containing the target face; projecting the target face 3D texture image to a 2D plane to obtain a target face 2D image; extracting feature information from the target face 2D image; comparing the feature information of the target face 2D image with feature information of a reference face 2D image to determine a similarity value; and in response to that the similarity value exceeds a first threshold, determining that the identity authentication succeeds.