This repository contains the code and resources for a research project focused on classifying emotional states using EEG signals. The study explores the performance of three advanced deep learning models: ShallowFBCSPNet, Deep4Net, and EEGNetv4. These models are specifically designed to process EEG data and classify emotions such as happiness, sadness, disgust, neutrality, and fear.
- Implementation of three neural network architectures optimized for EEG data.
- Preprocessing of raw EEG signals, including filtering, segmentation, and feature extraction.
- Training and evaluation of models using the SEED-V dataset.
- Analysis of model performance with visualizations such as confusion matrices and loss curves.
To run the code in this repository, you will need the following software and libraries:
- Python 3.10 or higher
- Google Colab (recommended for running with GPU support)
- MNE-Python for EEG signal processing
- PyTorch for implementing and training deep learning models
- Matplotlib for visualizing results
- NumPy and Pandas for data manipulation