[Review on identity feature extraction methods based on electroencephalogram signals].

Area of Science:

• Neuroscience
• Computer Science
• Biometrics

Background:

• Biometrics are crucial for information security.
• Electroencephalogram (EEG) signals represent a novel biometric modality with inherent advantages like safety and durability.
• EEG-based individual identification is a rapidly growing research area.

Purpose of the Study:

• To review and categorize existing identity feature extraction methods for EEG signals.
• To elucidate the principles, applications, and achievements of various EEG feature extraction techniques.
• To identify current challenges and forecast future research directions in EEG-based biometrics.

Main Methods:

• Review of single-channel EEG features.
• Analysis of inter-channel EEG features.
• Exploration of deep learning approaches for EEG feature extraction.
• Examination of spatial filter-based feature extraction methods.

Main Results:

• Various feature extraction methods, including single-channel, inter-channel, deep learning, and spatial filtering, are effective for EEG-based identification.
• Each method offers distinct advantages and applications depending on the specific requirements.
• The field is advancing with sophisticated techniques for extracting discriminative information from EEG data.

Conclusions:

• Effective feature extraction is paramount for high-performance EEG biometrics.
• Continued research into novel feature extraction methods is essential for advancing the field.
• Future trends point towards more sophisticated and integrated approaches for robust individual identification using EEG.