Face Authentication via Dynamic Visual Target Alignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing face authentication methods are vulnerable to fraud through the presentation of videos mimicking imposed movements, lacking robustness in detecting such fraudulent activities.

Innovation Solution

A method involving a terminal with an imager and screen that determines the pose of a face, displays a visual cue and target, and requires the user to align the cue with the target, incorporating pose estimation and photogrammetric reconstruction to authenticate the face, ensuring the alignment is maintained over time and the gaze is directed towards the target.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If challenge-based authentication methods are used to detect fraud, then authentication reliability is improved, but the system becomes vulnerable to video-based fraud

Engineering Contradiction:
Improveauthentication reliabilityVSAvoidvideo-based fraud vulnerability
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent implements a dynamic target position that changes between authentication attempts and requires continuous real-time tracking during the authentication process. The target position is updated based on the user's head pose estimation, creating a dynamic challenge that cannot be pre-recorded in a video. This dynamic element directly counters video-based fraud while maintaining authentication reliability.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system continuously estimates head pose from the camera feed and uses this feedback to update the target position in real-time. The user must actively adjust their head position to keep the orientation cue aligned with the moving target, creating a closed-loop feedback system that prevents video replay fraud while ensuring reliable authentication.

Inventive Principle:
Principle #23Feedback

2Reliability

If multiple authentication parameters are verified to reduce fraud, then authentication robustness is improved, but device complexity increases

Engineering Contradiction:
Improveauthentication robustnessVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent uses a single camera to perform multiple functions: capturing the user's face, estimating head pose, tracking eye gaze direction, and verifying authentication. By making the camera a multi-functional component that handles both imaging and pose estimation, the system achieves robust multi-parameter verification without proportionally increasing device complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system uses the existing camera hardware to automatically estimate head pose and track gaze direction through image processing algorithms. Rather than adding separate sensors or devices for each authentication parameter, the system makes the camera self-sufficient by extracting multiple authentication-relevant features (face position, head orientation, eye gaze) from the same image stream, thereby reducing overall system complexity.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3522053B1Method for authenticating a face
Publication Date: 2021.03.31 IDEMIA IDENTITY & SECURITY FRANCE SAS
  • EP3522053B1 patent drawingFigure 1
  • EP3522053B1 patent drawingFigure 2~3
  • EP3522053B1 patent drawingFigure 4a~5b

AI summary

The invention relates to a method for authenticating a face presented to a terminal comprising the steps in which: - the imager acquires an initial image of a face, - the processing unit determines a pose of the face from the initial image, and determines a reference pose and a target position placed randomly or pseudo-randomly in a target space, - the screen displays a displayed image (6) comprising at least one visual orientation marker (7) initially at a reference position, and a visual target (8) at the target position, updated by moving the visual orientation marker (7) according to the successive poses of the face, - the processing unit authenticates the presented face if the position of the visual orientation marker (7) coincides with the target position where the visual target (8) is located.