Single trial EEG classification applied to a face recognition experiment using different feature extraction methods.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Brain-machine interfaces (BMIs) are rapidly advancing.
• Electroencephalogram (EEG) classification is crucial for BMIs.
• Few studies explore EEG-based BMI for human face cognition.

Purpose of the Study:

• To evaluate feature extraction methods for EEG-based face recognition.
• To compare common spatial pattern, PCA, wavelet transform, and interval features.
• To assess BMI system performance in distinguishing familiar from novel faces.

Main Methods:

• Collected high-resolution EEG data from 15 healthy subjects.
• Presented subjects with familiar and novel faces.
• Implemented and compared four feature extraction techniques: common spatial pattern, principal component analysis, wavelet transform, and interval features.

Main Results:

• Principal component analysis (PCA) demonstrated superior performance.
• PCA achieved an average classification accuracy of 94.2%.
• Other methods showed lower classification accuracies compared to PCA.

Conclusions:

• PCA is a highly effective method for EEG-based face recognition.
• This research contributes to developing BMI systems for face recognition.
• Findings suggest potential real-life applications for EEG-based face familiarity detection.