Face Liveness Detection via 3D Pose Homography Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Facial recognition systems are vulnerable to spoofing attacks, particularly due to the need for multiple images with varying 3D poses to differentiate between real 3D faces and 2D facial images, which affects usability and turnaround time.
Innovation Solution
A method and system for face liveness detection that uses a single image to determine whether a face is real 3D or 2D by calculating homography mapping between 2D and 3D facial landmark points, eliminating the need for multiple image captures and reducing the complexity of sensor calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple images with varying 3D poses are used to detect presentation attacks, then the reliability of facial recognition is improved, but the loss of time and device complexity increase
Solution Approach 1:
The system performs sensor calibration and 3D shape capture during the registration phase beforehand, so that during the recognition phase, only a single image needs to be processed. The pre-acquired 3D shape information and calibrated sensor parameters are stored and reused, eliminating the need for repeated calibration and multiple image captures, thus reducing turnaround time while maintaining detection accuracy
Solution Approach 2:
The patent transitions from 2D image analysis to 3D shape-based analysis by capturing and utilizing the 3D geometry of the face. This dimensional enhancement allows the system to detect presentation attacks more effectively using a single image, as the 3D structure provides additional information that distinguishes real faces from spoofing attempts without requiring multiple 2D images from different poses
2Reliability
If multiple images with varying 3D poses are used to detect presentation attacks, then the reliability of facial recognition is improved, but the device complexity increases
Solution Approach 1:
The system performs sensor calibration and 3D shape capture during the registration phase beforehand, so that during the recognition phase, only a single image needs to be processed. The pre-acquired 3D shape information and calibrated sensor parameters are stored and reused, eliminating the need for repeated calibration and multiple image captures, thus reducing turnaround time while maintaining detection accuracy
Solution Approach 2:
The system creates a 3D digital model (copy) of the user's face during registration, which is then used for comparison during recognition. This 3D model serves as a reference that can be repeatedly compared against new images without requiring repeated complex calibration procedures, simplifying the recognition process while maintaining high detection reliability
3Measurement precision
If 3D shape capture is performed during recognition phase, then the detection precision is improved, but the loss of substance and productivity decrease
Solution Approach 1:
The system performs sensor calibration and 3D shape capture during the registration phase beforehand, so that during the recognition phase, only a single image needs to be processed. The pre-acquired 3D shape information and calibrated sensor parameters are stored and reused, eliminating the need for repeated calibration and multiple image captures, thus reducing turnaround time while maintaining detection accuracy
Data Source
Figure 1~3
Figure 4
Figure 5
AI summary
One embodiment provides a method for face liveness detection. The method comprises receiving a first image comprising a face of a user, determining one or more two-dimensional (2D) facial landmark points based on the first image, and determining a three-dimensional (3D) pose of the face in the first image based on the one or more determined 2D facial landmark points and one or more corresponding 3D facial landmark points in a 3D face model for the user. The method further comprises determining a homography mapping between the one or more determined 2D facial landmark points and one or more corresponding 3D facial landmark points that are perspectively projected based on the 3D pose, and determining liveness of the face in the first image based on the homography mapping.