Learning Domain- and Class-Disentangled Prototypes for Domain-Generalized EEG Emotion Recognition

• A Pytorch implementation of our under reviewed paper "Learning Domain- and Class-Disentangled Prototypes for Domain-Generalized EEG Emotion Recognition".

Installation

• Python 3.8
• Pytorch 2.0.0
• NVIDIA CUDA 11.8
• NVIDIA CUDNN 8700
• Numpy 1.24.3
• Scikit-learn 0.22.1
• scipy 1.5.2
• GPU NVIDA GeForce RTX 3090

Databases

• SEED, SEED-IV, SEED-V

Training and Testing

• Data Process Module: utils.py
• One-Hot processing: DataProcessing_OneHot.py
• Dynamically Updating Gradients: StepwiseLR_GRL.py
• MATL-DC Framework : MATL-DC framework.py
• Pairwise Learning Module: Pairwise_Learning_Modual.py
• MATL_DC_Train_verification: MATL_DC_Train.py

Usage

• After modify setting (path, etc), just run the main function in the MATL_DC_Train.py.