Structured Light Face Verification for Deepfake Video Calls

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video conferencing systems fail to continuously verify the identity of users during a call, allowing attackers to inject fake media streams, which undermines security and integrity.

Innovation Solution

Implement a system with a structured irradiation projector that generates a pseudo-random dot pattern, capturing a depth map of the user's face for continuous biometric verification, ensuring genuine real-time media streams by comparing with stored biometric data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If current video conferencing systems use only password and biometric data verification at login, then user authentication is simple and quick, but security is compromised allowing fake media stream injection during calls

Engineering Contradiction:
ImprovesecurityVSAvoidverification system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system performs preliminary biometric verification at login and then continuously performs micro-challenge verification throughout the video call. This preliminary action establishes baseline security while the continuous micro-challenges prevent fake media injection without requiring complete system redesign

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The verification process is segmented into discrete micro-challenges (e.g., blink detection, head rotation detection) that are independently verified throughout the call. This segmentation allows complex security verification to be broken into simple, manageable checks that maintain security without overwhelming system complexity

Inventive Principle:
Principle #1Segmentation

2Reliability

If continuous biometric verification using structured light and depth mapping is implemented, then deepfake detection capability is enhanced, but processing time and computational resources increase

Engineering Contradiction:
Improvedeepfake detection accuracyVSAvoidverification processing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

Instead of performing full biometric analysis continuously, the system performs partial verification through micro-challenges that require only specific facial actions (blink, head rotation). This partial action provides sufficient deepfake detection capability while significantly reducing processing time and computational overhead compared to continuous full biometric mapping

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system implements periodic micro-challenges at intervals during the video call rather than continuous verification. This periodic action maintains deepfake detection effectiveness while allowing normal video communication to proceed uninterrupted, minimizing perceived processing time and maintaining user experience

Inventive Principle:
Principle #19Periodic action

3Object-affected harmful factors

If standard security measures like user authentication and media encryption are used, then basic security is provided, but vulnerabilities remain at intermediary components where fake media can be introduced

Engineering Contradiction:
Improvevulnerability to fake media injectionVSAvoidsystem architecture complexity
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The system incorporates continuous feedback loops where biometric verification results directly control media stream transmission. If verification fails or depth map analysis detects inconsistencies, the system immediately blocks the media stream. This feedback mechanism closes the security loop, preventing fake media injection without requiring fundamentally different system architecture

Inventive Principle:
Principle #23Feedback

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

Prevents the injection of deepfake media streams by continuously authenticating users, maintaining security and integrity throughout the video call.

Implementation Method 1

capturing a reflected pattern from the user's face using an irradiation sensor to produce a depth map of the user's face

Methodology Applied
Scientific EffectLight reflection: Reflection

Data Source

PatentEP4685759A1Method, system and computer program product for providing deep fake detection and prevention in a video conference
Publication Date: 2026.01.28 PEXIP
  • EP4685759A1 patent drawingFigure 1
  • EP4685759A1 patent drawingFigure 2
  • EP4685759A1 patent drawingFigure 3

AI summary

A system and method for continuous biometric verification during video calls to prevent the injection of deepfake media streams. The method involves actively illuminating the user's face with near-infrared light and projecting a pseudo-random dot pattern onto the face using a structured light projector. An infrared camera or LiDAR/ToF sensor captures the reflected pattern to create a depth map, which is compared with stored biometric data for continuous identity verification. The system ensures secure transmission of genuine real-time media streams by continuously matching the depth map and video feed with the biometric database, detecting and blocking any discrepancies. Key components include a dot pattern generator, structured light projector, infrared camera, LiDAR/ToF sensor, triangulation algorithm, biometric database, and face identification algorithm. The invention enhances security in digital communication platforms by providing robust protection against fraudulent activities during video calls.