Facial expression recognition based on Electroencephalogram and facial landmark localization

Area of Science:

• Neuroscience
• Computer Science
• Biomedical Engineering

Background:

• Facial expression recognition is crucial for affective computing, mental illness diagnosis, and rehabilitation.
• Accurate facial expression recognition is a growing area of research with significant implications.

Purpose of the Study:

• To improve the accuracy of facial expression recognition using Electroencephalogram (EEG) signals.
• To develop a novel fusion method combining EEG and facial landmark data.

Main Methods:

• EEG signals were preprocessed using discrete wavelet transform (DWT) to extract energy features.
• Facial landmark localization was performed on facial images to extract relevant facial features.
• A fusion feature vector was created by combining EEG and facial features, then classified using a Support Vector Machine (SVM).

Main Results:

• The proposed fusion method achieved an accuracy of 86.94 ± 4.35%.
• This represents a 4.16% accuracy increase compared to EEG-based recognition (82.78 ± 5.78%).

Conclusions:

• The fusion method offers higher accuracy for facial expression recognition.
• The proposed approach demonstrates a stronger generalization capability.