GitHub - tomazas/itc2017: Code for paper "Application of Convolutional Neural Networks to Four-Class Motor Imagery Classification Problem"

About

"Application of Convolutional Neural Networks to Four-Class Motor Imagery Classification Problem"
by Tomas Uktveris and Vacius Jusas

MATLAB 2016a or later w/ CUDA capable GPU.

Download BCI IV 2008 EEG preprocessed data from: https://github.com/tomazas/icist2015/tree/master/data
Update path in config.m to point to the EEG data on disk
Launch an experiment in MATLAB: run experiment_cnn_xxxx.m

N.B. Some experiments take a long time to complete!

Check ./cnn_results directory.

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
2dmaps		2dmaps
ITC-paper		ITC-paper
classifiers		classifiers
cnn_results		cnn_results
data		data
features		features
utils		utils
vlfeat		vlfeat
LICENSE		LICENSE
README.md		README.md
cnn_arch.log		cnn_arch.log
cnn_batch.log		cnn_batch.log
cnn_bilinear.log		cnn_bilinear.log
cnn_epoch.log		cnn_epoch.log
cnn_filters.log		cnn_filters.log
cnn_imgsize.log		cnn_imgsize.log
cnn_learn_rate.log		cnn_learn_rate.log
cnn_momentum.log		cnn_momentum.log
cnn_xfold.log		cnn_xfold.log
config.m		config.m
experiment_cnn_arch.m		experiment_cnn_arch.m
experiment_cnn_batch_size.m		experiment_cnn_batch_size.m
experiment_cnn_epoch.m		experiment_cnn_epoch.m
experiment_cnn_filters.m		experiment_cnn_filters.m
experiment_cnn_imgsize.m		experiment_cnn_imgsize.m
experiment_cnn_learn_rate.m		experiment_cnn_learn_rate.m
experiment_cnn_momentum.m		experiment_cnn_momentum.m
experiment_cnn_xfold_all.m		experiment_cnn_xfold_all.m
experiment_cnn_xfold_bilinear.m		experiment_cnn_xfold_bilinear.m
log4m.log		log4m.log
log4m.m		log4m.m
preprocess.m		preprocess.m