/DRL

Code for paper "Decoupling Representation Learning for Imbalanced Electroencephalography Classification in Rapid Serial Visual Presentation Task”

Primary LanguagePython

Introduction

This is the implementation for Decoupling Representation Learning for Imbalanced Electroencephalography Classification in Rapid Serial Visual Presentation Task in Journal of Neural Engineering.

DRL.png

License

This project is released under the Apache 2.0 license.

Paradigm

paradigm

Installation

python >= 3.6
torch >= 1.7.0
numpy >= 1.20
tqdm >= 4.59.0
scipy >= 1.6.2

Usage

  1. modified the config file

    {
    "dataset":  
    {
        "name": "Public",  
        "subject_num": 1,
        "pairNum": 20000,
        "channel": 64
    },
    "train_para":
    {
        "batchsize_stage_1" : 1024,
        "batchsize_stage_2" : 128,
        "epoch_stage_1" : 60,
        "epoch_stage_2" : 20
    },
    "model_para":
    {
        "F1": 8, 
        "F2": 2,
        "D": 1,
        "kernel_size": 3,
        "droup_out": 0.6
    }
}

  2. Train model

    python main.py

Citation

If you use this toolbox or benchmark in your research, please cite this project.

@article{Li_2022,
doi = {10.1088/1741-2552/ac6a7d},
url = {https://doi.org/10.1088/1741-2552/ac6a7d},
year = 2022,
month = {may},
publisher = {{IOP} Publishing},
volume = {19},
number = {3},
pages = {036011},
author = {Fu Li and Hongxin Li and Yang Li and Hao Wu and Boxun Fu and Youshuo Ji and Chong Wang and Guangming Shi},
title = {Decoupling representation learning for imbalanced electroencephalography classification in rapid serial visual presentation task},
journal = {Journal of Neural Engineering}}