Published in Multimedia Tools and Applications, 2025
ECG-based biometric ID using EMD+GRU achieves up to 99.89% accuracy on public datasets, outperforming SOTA methods.
Recommended citation: Zehir, H., Hafs, T., Daas, S., & Nait-ali, A. (2025). EMD based biometric identification system from electrocardiogram signals using GRU neural networks. Multimedia Tools and Applications, 1-20.
Download Paper
Published in Arabian Journal for Science and Engineering, 2025
ECG+voice multimodal biometrics using CNNs achieves 100% accuracy via score-level fusion, boosting security and robustness in identification.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. (2025). Unifying Heartbeats and Vocal Waves: An Approach to Multimodal Biometric Identification At the Score Level. Arabian Journal for Science and Engineering, 1-20.
Download Paper
Published in International Journal of Signal and Imaging Systems Engineering, 2025
Optimized quantized CNN for ECG biometrics hits 97.9% accuracy, 77% faster and 64% smaller, enabling efficient ESP32-based identification.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. (2025). Hardware-Optimised CNN Architecture for ECG Biometric Identification on Embedded Systems. International Journal of Signal and Imaging Systems Engineering
Published in Evolving Systems, 2024
ECG biometrics using EMD with GRU/LSTM models achieves up to 99.17% accuracy across three databases, showing strong potential for secure identification.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. (2024). Empirical mode decomposition-based biometric identification using GRU and LSTM deep neural networks on ECG signals. Evolving Systems, 15(6), 2193-2209.
Download Paper
Published in International Journal of Signal and Imaging Systems Engineering, 2024
ECG spectrograms classified by INNs achieve up to 97.93% accuracy, outperforming CNNs by better capturing local and temporal heartbeat patterns.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. (2024). Involutional neural networks for ECG spectrogram classification and person identification. International Journal of Signal and Imaging Systems Engineering, 13(1), 41-53.
Published in Romanian Journal of Information Science and Technology, 2024
Score-normalized multimodal biometrics with EMD achieves 1.69% EER by fusing online signature and fingerprint, boosting accuracy and fraud resistance.
Recommended citation: Hafs, T., Zehir, H., Hafs, A., Brahmia, H., & Nait-Ali, A. (2024). Enhancing Recognition in Multimodal Biometric Systems: Score Normalization and Fusion of Online Signatures and Fingerprints. SCIENCE AND TECHNOLOGY, 27(1), 37-49.
Download Paper
Published in Applied Computer Systems, 2023
Score-normalized multimodal biometrics that use EMD to combine a signature and a fingerprint, making it more accurate and less likely to be forged.
Recommended citation: Hafs, T., Zehir, H., Hafs, A., & Nait-Ali, A. (2023). Multimodal Biometric System Based on the Fusion in Score of Fingerprint and Online Handwritten Signature. Appl. Comput. Syst., 28(1), 58-65.
Published in Journal of Engineering Studies and Research, 2023
ECG biometrics using spectral features and SVMs achieves up to 97.0% accuracy on MIT-BIH, offering a secure and forgery-resistant identification method.
Recommended citation: Zehir, H., Hafs, T., Daas, S., & Nait-Ali, A. (2023). Support vector machine for human identification based on non-fiducial features of the ecg. Journal of Engineering Studies and Research, 29(1), 61-69.
Download Paper
Published in The 1st International Conference on Electrical Engineering & Renewable Energies Systems, 2024
Optimized quantized CNN for speaker recognition on ESP32 hits 98.58% accuracy, 53% faster and 73% smaller, enabling efficient edge deployment.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. TinyCNN: An Embedded CNN Model for Speaker Identification Using ESP32.
Download Paper | Download Slides
Published in ICAECCT23: The 1st International Conference on Advances in Electronics, Control and Computer Technologies, 2024
TinyML-based ECG biometric system on ESP32 achieves 96.71% accuracy, enabling resource-efficient personal identification.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. ECG-Based Biometric System using TinyML: Implementation and Performance Evaluation on ESP32.
Download Paper | Download Slides
Published in 2023 International Conference on Decision Aid Sciences and Applications, 2024
ECG QRS-based biometrics achieve 99.47% accuracy on healthy subjects and 97.60% on mixed cases, offering secure, reliable identification for healthcare and security.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. ECG-Based Biometric System using TinyML: Implementation and Performance Evaluation on ESP32.
Download Paper | Download Slides
Published in 4th International Conference on Technological Advances in Electrical Engineering, 2024
Edge-based TinyML online signature ID on ESP32 achieves 94.94% accuracy, proving effective for secure, real-time identification on resource-limited devices.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. Edge Based Online Signature Identification: A TinyML Approach with ESP32 Microcontroller.
Download Paper
Published in 2023 5th International Conference on Bio-engineering for Smart Technologies, 2023
ECG biometrics using HHT-derived instantaneous frequencies and GRU models achieve 96.42% (PTB) and 95.31% (MIT-BIH), proving robust identification performance.
Recommended citation: Zehir, H., Hafs, T., Daas, S., & Nait-Ali, A. (2023, June). An ecg biometric system based on empirical mode decomposition and hilbert-huang transform for improved feature extraction. In 2023 5th International Conference on Bio-engineering for Smart Technologies (BioSMART) (pp. 1-4). IEEE.
Download Paper
Published in 5th International Conference on Embedded Systems in Telecommunications and Instrumentation, 2022
ECG biometrics using EEMD+HHT features and BiLSTM achieve 97.42% accuracy (5 subjects) and 89.33% (full PTB dataset), confirming feature effectiveness.
Recommended citation: Zehir, H., Hafs, T., & Daas, S. Bidirectional Long Short-term Memory Neural Networks Based Electrocardiogram Biometric System.
Download Paper