A Novel Anti-Spoofing Solution for Iris Recognition Toward Cosmetic Contact Lens Attack Using Spectral ICA Analysis

Area of Science:

• Biometrics
• Computer Vision
• Image Processing

Background:

• Iris recognition is a key biometric technology.
• Cosmetic contact lenses (CCLs) can interfere with iris recognition systems.
• Distinguishing natural irides and CCL patterns is crucial for accurate identification.

Purpose of the Study:

• To develop an anti-spoofing method for iris recognition against CCLs.
• To enhance the accuracy of iris pattern recognition in the presence of CCLs.
• To restore genuine iris patterns obscured by CCLs.

Main Methods:

• Utilized a dual-band spectral imaging system to capture iridal images.
• Employed independent component analysis (ICA) to separate spectral primitives.
• Developed an ICA-based anti-spoofing scheme to differentiate iris and CCL patterns.

Main Results:

• Successfully distinguished natural iris texture from CCL patterns.
• Restored genuine iris patterns from CCL-polluted images.
• Improved recognition accuracy, reducing the False Rejection Rate (FRR) from 10.52% to 0.57% in a proof-of-concept study with 200 image pairs.

Conclusions:

• The proposed ICA anti-spoofing scheme is effective in mitigating CCL interference in iris recognition.
• The method significantly enhances the reliability and accuracy of iris-based biometric systems.
• This approach offers a robust solution for secure and accurate iris recognition.