Signal Quality and Cross-Session Robustness in ECG Biometric Authentication: A Review of Methods and Challenges

ECG-based biometric authentication depends on effective signal preprocessing, reliable quality assessment, and appropriate verification protocols. The reviewed literature highlights the importance of distinguishing identification from verification and evaluating systems using metrics such as FAR, FRR, and EER. Signal quality and cross-session variations remain important challenges that can affect authentication reliability. Quality-based rejection may improve the reliability of accepted signals but can also reduce system coverage. Fair evaluation requires proper data partitioning, clear reporting of experimental settings, and prevention of data leakage. Furthermore, differences in datasets and acquisition conditions limit the generalizability of existing findings. Future research should explore advanced computational approaches for ECG feature learning and authentication, while addressing security, privacy, and robustness under varying recording conditions, to support the development of reliable and practical ECG-based biometric systems.These considerations can support the development of more reliable and practical ECG-based biometric authentication systems.

[1] “ECG-Based Biometric Recognition: A Survey of Methods and Databases,” Sensors, vol. 25, no. 6, Art. no. 1864, 2025, doi: 10.3390/s25061864. [2] “Biometric Recognition: A Systematic Review on Electrocardiogram Data Acquisition Methods,” Sensors, vol. 23, no. 3, Art. no. 1507, 2023, doi: 10.3390/s23031507. [3] “ECG Signal Feature Extraction Trends in Methods and Applications,” BioMedical Engineering OnLine, vol. 22, Art. no. 22, 2023, doi: 10.1186/s12938-023-01075-1. [4] P. Melzi, R. Tolosana, and R. Vera-Rodriguez, “ECG Biometric Recognition: Review, System Proposal, and Benchmark Evaluation,” IEEE Access, vol. 11, pp. 15555–15566, 2023, doi: 10.1109/ACCESS.2023.3244651. [5] T. S. Lugovaya, “Biometric Human Identification Based on Electrocardiogram,” M.S. thesis, Electrotechnical Univ. LETI, St. Petersburg, Russia, 2005. ECG-ID database, PhysioNet. [6] A. R. J. Mitchell, D. Ahlert, C. Brown, M. Birge, and A. Gibbs, “Electrocardiogram-based biometrics for user identification—Using your heartbeat as a digital key,” Journal of Electrocardiology, vol. 80, pp. 1–6, 2023, doi: 10.1016/j.jelectrocard.2023.04.001. [7] “Deep Learning-Powered ECG-Based Biometric Authentication,” in Proc. 2023 Int. Conf. Next-Generation Computing, IoT and Machine Learning (NCIM), 2023, doi: 10.1109/NCIM59001.2023.10212845.