Electroencephalogram-Based Facial Gesture Recognition Using Self-Organizing Map.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Human-Computer Interaction

Background:

• Brain-computer interfaces (BCIs) enable direct brain-to-computer communication, crucial for assistive technologies.
• Electroencephalograms (EEGs) are used in BCIs to interpret brain activity for environmental interaction.
• Controlling assistive devices like wheelchairs and prosthetic limbs is a key application for BCI systems.

Purpose of the Study:

• To develop and evaluate an electroencephalogram (EEG)-based facial gesture recognition system.
• To enhance human-computer interaction capabilities for individuals with disabilities through BCI technology.
• To investigate the effectiveness of a self-organizing map (SOM) approach for classifying EEG signals related to facial gestures.

Main Methods:

• Utilized EEG signals, specifically α, β, and θ power bands, as features for facial gesture recognition.
• Employed a Self-Organizing Map (SOM)-Hebb classifier for feature vector classification.
• Developed an online facial gesture recognition system using MATLAB, integrating facial movements detectable in EEG signals.

Main Results:

• The EEG-based facial gesture recognition system achieved accuracies ranging from 76.90% to 97.57%.
• The lowest accuracy (76.90%) was observed when recognizing seven distinct gestures.
• The developed online system demonstrated robust performance compared to existing EEG-based recognition methods, with a recognition flow time of 5.7 seconds.

Conclusions:

• The proposed EEG-based facial gesture recognition method using SOM is effective for BCI applications.
• The system offers a viable solution for controlling assistive devices, improving the quality of life for people with disabilities.
• Further research can explore expanding the number of recognizable gestures to enhance BCI system versatility.