Iris Biometric Recognition via 3D Geometric Verification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing biometric iris recognition methods are vulnerable to falsification by cosmetic lenses that mask or replicate iris patterns, leading to unreliable authentication.

Innovation Solution

A method involving capturing iris images from different orientations to detect homologous characteristic points, applying a homography to align them, and determining residual differences to verify the authenticity of the iris, distinguishing between real and replicated patterns by assessing the shape consistency of these points.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If cosmetic lenses with replicated iris patterns are used, then the appearance of the iris is modified, but the reliability of biometric recognition deteriorates

Engineering Contradiction:
Improveappearance modificationVSAvoidbiometric recognition reliability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent transitions from 2D iris pattern analysis to 3D geometric analysis by capturing images at multiple orientations. It extracts 3D coordinates of characteristic points and analyzes their spatial relationships, adding a dimensional aspect that reveals the curvature difference between real irises and flat lens replicas.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent replaces traditional optical/electrical iris recognition systems with a geometric verification system. Instead of analyzing iris texture patterns, it uses geometric relationships of characteristic points in 3D space, substituting the recognition mechanism from pattern-matching to spatial-relationship verification.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If multiple images from different orientations are captured and analyzed, then the ability to detect cosmetic lenses improves, but the device complexity increases

Engineering Contradiction:
Improvecosmetic lens detection accuracyVSAvoidimage capture and processing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system uses the geometric relationships of characteristic points themselves to verify authenticity. The characteristic points' spatial configurations inherently contain the verification information, eliminating the need for external reference data or complex authentication protocols. The iris structure verifies itself through its geometric properties.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent changes the verification parameter from 2D pattern matching to 3D geometric relationships. By capturing images at different orientations and analyzing the spatial coordinates of characteristic points, it transforms the verification approach to use dimensional and positional parameters instead of texture patterns.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If a homography is applied to align characteristic points from different orientations, then the measurement precision of geometric verification improves, but the computational requirements increase

Engineering Contradiction:
Improvecharacteristic points alignment accuracyVSAvoidcomputational processing power
Core Design Contradiction:
Measurement precisionVSPower

Solution Approach 1:

The patent performs homography transformation as a preliminary step before conducting the actual geometric verification. By pre-aligning the characteristic points from different orientations, it simplifies subsequent calculations and enables more efficient verification of the 3D geometric relationships.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3567521B1Iris biometric recognition method
Publication Date: 2023.08.23 IDEMIA IDENTITY & SECURITY FRANCE SAS

AI summary

A method for verifying the authenticity of an iris from an eye of a candidate for biometric recognition, comprising the step of verifying the flatness of the iris from two images of the iris taken from different orientations relative to the image sensor. A method and optical recognition device for implementing this method.