Hand biometric recognition based on fused hand geometry and vascular patterns

Area of Science:

• Biometrics and Security Engineering
• Computer Vision and Pattern Recognition
• Human-Computer Interaction

Background:

• Traditional biometric systems often rely on single modalities, which can be vulnerable to spoofing and environmental variations.
• Integrating multiple biometric traits (multimodal biometrics) enhances system robustness and accuracy.
• Existing hand-based biometrics may require multiple images or complex feature extraction processes.

Purpose of the Study:

• To propose a novel multimodal biometric authentication system integrating hand geometry and vascular patterns.
• To develop a cost-effective system capable of extracting features from a single image.
• To evaluate the performance of the proposed system in terms of accuracy and error rates.

Main Methods:

• Hand geometry features extracted include side-view thickness, K-curvature, finger valley dimensions, and finger profiles.
• Vascular patterns are acquired using a direction-based extraction method from the same single image.
• A score-level fusion approach combines the scores from hand geometry and vascular pattern recognition.

Main Results:

• The proposed multimodal biometric system successfully integrates hand geometry (side and back views) and vascular patterns.
• Feature extraction is efficiently performed using only a single image.
• The system demonstrated a remarkably low equal error rate (EER) of 0.06%.

Conclusions:

• The proposed multimodal biometric approach offers a robust and accurate authentication solution.
• The system's ability to use a single image and its potential for low-cost implementation make it practical for widespread adoption.
• The high accuracy (0.06% EER) validates the effectiveness of combining hand geometry and vascular patterns for biometric security.