Optical Skin Detection for 3D Face Unlock Spoof Defense
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition systems are vulnerable to spoof attacks using 3D masks and fail to provide reliable authentication across different skin types, leading to security breaches and computational inefficiencies.
Innovation Solution
A method for face authentication that includes face detection, skin detection using illumination patterns, and 3D detection to differentiate between real human faces and spoofing materials, utilizing a skin classification model and depth analysis to authenticate faces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If 3D cameras and deep convolutional neuronal networks are used for face recognition, then authentication accuracy is improved, but computational power requirements and processing time increase
Solution Approach 1:
The authentication process is divided into multiple independent stages: initial face detection using 2D image processing, skin detection using illumination patterns and color analysis, and 3D verification using depth maps. This segmentation allows the system to perform basic filtering with computationally efficient methods before applying more resource-intensive 3D analysis only when necessary
Solution Approach 2:
The system performs skin detection using color space analysis and illumination pattern matching as a preliminary filter before full 3D verification. This partial action catches many spoofing attempts early in the process, allowing the computationally expensive deep convolutional neuronal networks to be applied only to cases that pass the initial filters, reducing overall computational burden
2Reliability
If multiple video frames are processed for presentation attack detection, then detection reliability is improved, but processing speed decreases
Solution Approach 1:
The system performs skin detection using color analysis and illumination pattern matching before processing multiple video frames for 3D verification. This preliminary action identifies and filters out obvious spoofing attempts (such as photographs and videos) early in the process, so that full multi-frame processing is only applied to cases that pass the initial skin detection filter, maintaining both accuracy and speed
3Productivity
If expensive hardware is used to deliver acceptable unlock performance, then authentication speed is improved, but device cost and power consumption increase
Solution Approach 1:
The authentication pipeline is segmented into multiple processing stages with increasing computational requirements: 2D face detection, skin detection using color space analysis, illumination pattern verification, and 3D depth map analysis. This segmentation allows the system to use simpler, less expensive hardware components for the majority of the authentication process, reserving expensive high-performance hardware only for the final verification stage when needed
Solution Approach 2:
The system replaces computationally intensive mechanical processing with optical and color-based detection methods. By using illumination patterns and color space analysis to detect skin properties, the system achieves reliable spoofing detection without requiring expensive high-speed processing hardware, thereby reducing device cost and power consumption while maintaining authentication speed
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
Enhances security by reliably distinguishing between real faces and spoofing materials, operates across diverse skin types, and reduces computational demands, enabling fast and efficient authentication.
Implementation Method 1
determining at least one material property of the reflection feature from the first beam profile information... by using a skin classification model which is parametrized to take into account differences in the reflection of different zones of a face
Data Source
Figure 1~2
Figure 3
AI summary
A method for face authentication is proposed. The method comprising the following steps: a) at least one face detection step (110), wherein the face detection step (110) comprises capturing at least one first image by using at least one camera (112), wherein the first image comprises at least one two-dimensional image of a scene suspected to comprise the face, wherein the face detection step (110) comprises detecting the face in the first image by identifying in the first image at least one predefined or pre-determined geometrical feature characteristic for faces by using at least one processing unit (114); b) at least one skin detection step (116), wherein the skin detection step (116) comprises projecting at least one illumination pattern comprising a plurality of illumination features on the scene by using at least one illumination unit (118) and capturing at least one second image using the at least one camera (112), wherein the second image comprises a plurality of reflection features generated by the scene in response to illumination by the illumination features, wherein each of the reflection features comprises at least one beam profile, wherein the skin detection step comprises determining a first beam profile information of at least one of the reflection features located inside an image region of the second image corresponding to an image region of the first image comprising the identified geometrical feature by analysis of its beam profile and determining at least one material property of the reflection feature from the first beam profile information by using the processing unit (114), wherein the detected face is characterized as skin if the material property corresponds to at least one property characteristic for skin; c) at least one 3D detection step (120), wherein the 3D detection step (120) comprises determining a depth map of at least parts of the scene by determining at least one depth information of the reflection features located inside the image region of the second image corresponding to the image region of the first image comprising the identified geometrical feature, wherein the 3D detection step (120) comprises using 3D topology data of the face in front of the camera (112); d) at least one authentication step (122), wherein the authentication step (122) comprises authenticating the detected face by using at least one authentication unit if in step b) (116) the detected face is characterized as skin and in step c) (120) the detected face is characterized as 3D object.