Hewlbern / EnK-research

EnK: Encoding time-information in convolution

EnK: Encoding time-information in convolution

Singh, A.K. and Lin, C.T., 2020. EnK: Encoding time-information in convolution. arXiv preprint arXiv:2006.04198.

Requirements

• Python == 3.7 or 3.8
• tensorflow == 2.X (both for CPU and GPU)
• PyRiemann >= 0.2.5
• scikit-learn >= 0.20.1
• matplotlib >= 2.2.3

How to run

• Input Data Format: Number of EEG Channels x Number of Samples X Number of Trials for EEG data and Labels as a vector. See testData.mat for references with sampling rate of 400 Hz.
• Provide input data related information in 'op.py' such as path, sampling rate, number of classes, etc.
• Execute the following line of code

python main.py

Models implemented/used

• EEGNet [1]
• DeepConvNet [2]
• ShallowConvNet [3]

EEGNet, DeepConvNet, and ShallowConvNet implementation are based on EEGNet repo [4]