Introduction

This repository provides source code for replication of SSVEPNet [1]. SSVEPNet is an efficient CNN-LSTM mixed network with spectral normalization and label smoothing technologies. It aims to reduce the gap of classification performance between intra- and inter-subject experiments and implement a high-performance SSVEP-BCI system with limited training data.

File Description

• data/Dial: A 12-class public SSVEP dataset [2].
• Model/SSVEPNet.py: Implementation of CNN-LSTM mixed network with pytorch.
• Result/DatasetA: Recorded classification results.
• Train/Classifier_Trainer.py: Training code for any DL-based classifier model.
• Test/SSVEPNet_Test.py: Evaluating the performance for SSVEPNet.
• Utils/EEGDataset.py: Providing the data for intra- and inter- subject experiments.
• Utils/Constraint.py: Implementation of spectral normalization and spatial convolution kernel constraint.
• Utils/LossFunction.py: Implementation of visual attention based label smooth techniques.
• Utils/Ploter: Plotting histogram to visualize the classification results.

Running Environment

• Setup a virtual environment with python 3.8 or newer
• Install requirements

pip install -r Resource/requirements.txt

Train and Test

cd Test
python SSVEPNet_Test.py

Reference

[1] Pan Y, Chen J, Zhang Y, et al. An efficient CNN-LSTM network with spectral normalization and label smoothing technologies for SSVEP frequency recognition[J]. Journal of Neural Engineering, 2022, 19(5): 056014. https://iopscience.iop.org/article/10.1088/1741-2552/ac8dc5/meta

[2] Nakanishi M, Wang Y, Wang Y T, et al. A comparison study of canonical correlation analysis based methods for detecting steady-state visual evoked potentials[J]. PloS one, 2015, 10(10): e0140703. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0140703