Cooperative chameleon search optimization enabled deep ensemble classifier for emotion recognition using EEG signal.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Machine Learning

Background:

• Accurate emotion recognition is crucial for understanding behavioral and cognitive functioning.
• Existing Electroencephalography (EEG) signal analysis methods face challenges in feature extraction and subject dependency.
• Developing efficient emotion recognition techniques is essential for various applications.

Purpose of the Study:

• To propose an effective emotion recognition model using EEG signals.
• To address limitations in feature extraction and subject dependency in current methods.
• To enhance the accuracy and reliability of emotion detection from EEG data.

Main Methods:

• Developed a Cooperative Chameleon Search Optimization-enabled Deep Ensemble Convolutional Recurrent Neural network (CC-DERN) classifier.
• Utilized statistical features, Frequency Domain Features (FDF), Time-Domain Features (TDF), and VGG-16 for feature extraction.
• Implemented enhanced preprocessing techniques to remove artifacts from EEG signals.
• Employed Cooperative Chameleon Search Optimization for discriminative feature selection.

Main Results:

• The CC-DERN model achieved high performance metrics on both the DEAP and SEED datasets.
• DEAP dataset results: 96.09% accuracy, 99.551% sensitivity, and 93.430% specificity.
• SEED dataset results: 98% accuracy, 99.551% sensitivity, and 84.024% specificity.
• Demonstrated superior performance compared to traditional emotion recognition techniques.

Conclusions:

• The CC-DERN model offers a robust and accurate solution for emotion recognition from EEG signals.
• The integration of Cooperative Chameleon Search Optimization significantly improves feature selection and detection accuracy.
• The proposed method effectively overcomes challenges related to artifact removal and subject dependency in EEG-based emotion recognition.