Face Anti-Counterfeiting Detection Using Neural Networks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current face anti-counterfeiting detection methods rely on specialized hardware and multi-spectral devices, which are costly and limit their applicability, and struggle to differentiate between real and counterfeited faces effectively without interactive processes.

Innovation Solution

A face anti-counterfeiting detection method and system that utilize a neural network to extract features from images or videos using visible light cameras, identifying counterfeited face clue information such as edge, reflection, and material information without requiring special hardware, allowing for effective detection under visible light conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If specialized hardware and multi-spectral devices are used for face anti-counterfeiting detection, then detection precision is improved, but device complexity and cost increase

Engineering Contradiction:

Improvedetection precisionVSAvoiddevice complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent uses a visible light camera to capture images that copy or replicate the information that would otherwise require specialized multi-spectral devices. By training a neural network to recognize counterfeiting clues in visible light images, the system creates a computational copy of the detection capability that would otherwise require expensive hardware

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces the mechanical/optical system of specialized multi-spectral devices with a computational system using a visible light camera and neural network. The neural network algorithm substitutes for the physical multi-spectral sensing hardware, achieving similar detection functionality through software rather than complex hardware

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If specialized hardware and multi-spectral devices are used for face anti-counterfeiting detection, then detection precision is improved, but cost increases

Engineering Contradiction:

Improvedetection precisionVSAvoidcost

Core Design Contradiction:

Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent replaces expensive, specialized multi-spectral devices with inexpensive visible light cameras that are mass-produced and widely available. The neural network model serves as a software solution that can be deployed on standard hardware, dramatically reducing the cost of implementation while maintaining detection effectiveness

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Measurement precision

If interactive processes are used for face anti-counterfeiting detection, then detection precision is improved, but ease of operation worsens

Engineering Contradiction:

Improvedetection precisionVSAvoidease of operation

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent implements silent living body detection where the system automatically performs detection without requiring user interaction. The neural network analyzes the captured image or video frame directly to identify counterfeiting clues, eliminating the need for users to perform interactive actions while maintaining high detection precision

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS11482040B2Face anti-counterfeiting detection methods and systems, electronic devices, programs and media

Publication Date: 2022.10.25 BEIJING SENSETIME TECH DEV CO LTD

AI summary

A face anti-counterfeiting detection method includes: obtaining an image or video to be detected containing a face; extracting a feature of the image or video to be detected, and detecting whether the extracted feature contains counterfeited face clue information; and determining whether the face passes the face anti-counterfeiting detection according to a detection result.