A Review of ECG Biometrics: Generalization in Deep Learning with Attention Mechanisms

• الحاوية / القاعدة: Lecture Notes in Electrical Engineering
• المؤلف الرئيسي: 2-s2.0-85190385476
• التنسيق: Conference paper
• اللغة: English
• منشور في: Springer Science and Business Media Deutschland GmbH 2024
• الوصول للمادة أونلاين: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85190385476&doi=10.1007%2f978-981-99-9005-4_57&partnerID=40&md5=8707407e04af68edbb028ce329e321e3
• تدمد: 18761100
• DOI: 10.1007/978-981-99-9005-4_57

Common biometric systems like fingerprint and face recognition are more convenient in daily applications, but biological and behavioral characteristics of the biometrics features can be fabricated and digitally stolen. Thus, biometrics features with liveness detection such as the electrocardiogram (ECG) have been introduced as its features are hidden and difficult to forge. This study presents a review of ECG biometrics based on deep learning and generalization issues in deep learning. Based on the review, deep learning methods such as recurrent neural networks (RNN) and long short-term memory networks (LSTM) with attention mechanisms can be employed to improve the performance and generalization ability of ECG biometrics systems. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.