ECG Biometric Recognition Using Adaptive Bloom Filter-Based Template Construction.

Area of Science:

• Biometrics
• Signal Processing
• Cybersecurity

Background:

• Biometric recognition offers an alternative to memorized credentials.
• Electrocardiogram (ECG) biometrics show potential but face privacy and security challenges due to weak template protection.
• Existing ECG biometric systems struggle with safeguarding sensitive user data.

Purpose of the Study:

• To introduce novel Bloom filter-based strategies for generating secure biometric templates from ECG data.
• To enhance privacy and security in ECG-based identification and verification systems.
• To improve the robustness of ECG biometric systems against unauthorized access and variations.

Main Methods:

• Development of Bloom filter-based template generation techniques for ECG biometrics.
• Integration of nonlinear transformations to prevent template reconstruction.
• Implementation of an interquartile range-based method for identity confirmation.
• Evaluation of proposed schemes in both identification and verification modes.

Main Results:

• The proposed Bloom filter strategies effectively address privacy and security concerns.
• Nonlinear transformations significantly complicate template reconstruction.
• The interquartile range method enhances accuracy against variations and attacks.
• Achieved mean equal error rates of 7.9% (identification) and 1.3% (verification).

Conclusions:

• Bloom filter-based ECG biometric template generation offers a secure and effective solution.
• The proposed methods enhance user privacy and system security.
• The system demonstrates high accuracy and robustness in real-world scenarios.