An EEG-Based Identity Authentication System with Audiovisual Paradigm in IoT

Area of Science:

• Biometrics and Human-Computer Interaction
• Signal Processing and Machine Learning
• Cybersecurity and Internet of Things (IoT)

Background:

• Traditional identity authentication methods face limitations in security and trustworthiness for intelligent IoT applications.
• Electroencephalography (EEG) signals offer unique, stable, and universal characteristics suitable for biometric authentication.
• Existing EEG-based authentication systems suffer from limited experimental usability and suboptimal classification accuracy, hindering widespread adoption.

Purpose of the Study:

• To develop a more usable and accurate identity authentication system for intelligent IoT environments using EEG signals.
• To propose a novel audiovisual presentation paradigm for EEG signal acquisition.
• To enhance the performance of EEG-based identity authentication through advanced signal processing and machine learning techniques.

Main Methods:

• An audiovisual presentation paradigm was employed to record electroencephalography (EEG) signals.
• Artifact removal was performed using reference electrode, ensemble averaging, and independent component analysis.
• Feature extraction involved adaptive feature selection and bagging ensemble learning for optimal classification model development.

Main Results:

• The proposed audiovisual paradigm achieved superior classification accuracy compared to other existing paradigms and conventional EEG authentication methods.
• The system demonstrated feasibility, effectiveness, and reliability through testing in a practical login scenario.
• Advanced signal processing and ensemble learning significantly improved the performance of EEG-based identity authentication.

Conclusions:

• The novel audiovisual EEG authentication system offers a promising solution for secure and trustworthy identity verification in intelligent IoT applications.
• The proposed methodology effectively addresses the limitations of previous EEG-based authentication systems, particularly in terms of accuracy and usability.
• This research paves the way for the widespread implementation of EEG biometrics in real-world IoT security scenarios.