Facial expression classification using EEG and gyroscope signals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Facial expression recognition is crucial for human-computer interaction.
• Existing methods often require specialized hardware like electromyography (EMG) electrodes.
• Electroencephalography (EEG) signals contain muscle activity information from facial expressions.

Purpose of the Study:

• To classify six facial expressions using only signals from a low-cost EEG headset.
• To develop a hybrid brain-computer interface (BCI) system without additional hardware.
• To evaluate the impact of gyroscope data on facial expression classification accuracy.

Main Methods:

• Utilized muscle and gyroscope signals from a low-cost EEG headset.
• Extracted features from both time and frequency domains of EEG data.
• Employed optimized sampling rates and classification methods for each participant and feature set.
• Integrated gyroscope data with EEG features for classification.

Main Results:

• Achieved high accuracy in classifying six different facial expressions.
• Demonstrated that EEG signals capture muscle activities related to facial expressions.
• Showed an average performance increase of 9-16% by incorporating gyroscope data.
• Validated a novel BCI approach using existing EEG hardware.

Conclusions:

• Facial expression classification is feasible using low-cost EEG and embedded gyroscope data.
• A hybrid BCI system can be implemented without additional EMG hardware.
• Gyroscope integration significantly enhances the accuracy of facial expression recognition via BCI.