ResNet1D-Based Personal Identification with Multi-Session Surface Electromyography for Electronic Health Record
Raghavendra Ganiga1, Muralikrishna S N2, Wooyeol Choi3
1Department of Information and Communication Technology, Manipal Institute of Technology, Manipal Academy of Higher Education (MAHE), Manipal 576104, India.
Sensors (Basel, Switzerland)
|May 25, 2024
Summary
This study introduces a novel personal identification method using ResNet1D deep learning to analyze surface electromyography (sEMG) signals for secure electronic health record (EHR) access. The sEMG-based approach offers a potentially more secure alternative for safeguarding patient information.
Area of Science:
- Biometrics and Human-Computer Interaction
- Deep Learning and Signal Processing
- Health Informatics and Cybersecurity
Background:
- Traditional personal identification methods for electronic health records (EHRs) face security vulnerabilities and user inconvenience.
- Ensuring patient privacy and secure access to sensitive medical information is critical in healthcare systems.
- Existing authentication methods like passwords and biometrics can be compromised, necessitating advanced solutions.
Purpose of the Study:
- To develop and evaluate a novel personal identification system for EHRs using deep learning analysis of surface electromyography (sEMG) signals.
- To explore the potential of ResNet1D architecture for robust user authentication based on sEMG data.
- To offer a more secure and convenient alternative for accessing electronic health records.
Main Methods:
- Collected a multi-session sEMG signal database from 200 subjects performing hand gestures across three sessions.
- Utilized the ResNet1D deep learning model to analyze sEMG signals for discriminative feature extraction.
- Trained and validated the ResNet1D model for both gesture recognition and personal identification tasks within a simulated EHR system.
Main Results:
- The ResNet1D model achieved high identification accuracy, with rates of 97% for 5 subjects, 96% for 10 subjects, 87% for 15 subjects, and 82% for 20 subjects.
- The system demonstrated the ability to validate individual identities by comparing captured sEMG features against stored templates.
- Experimental results on a subset of the database confirmed the model's effectiveness in personal identification.
Conclusions:
- The proposed ResNet1D-based sEMG personal identification method offers a promising and secure alternative for EHR systems.
- This approach can significantly enhance the security and privacy of patient information within digital healthcare environments.
- Integration of this sEMG identification system into EHRs can lead to more reliable and protected access to medical data.
