Code to accompany our International Conference on Pattern Recognition (ICPR) paper entitled - Leveraging Synthetic Subject Invariant EEG Signals for Zero Calibration BCI.
Hi, i saw that when you create class EEG_CNN_Generaor , you firstly add a dense layer, i am wondering that why you use the value of 2816 in this full connection layer.
Well done. I am doing similar work these days. When i used GAN, I found that I couldn't generate the EEG data, only producing some noise signals. So i want you make sure if there is some error in the network like activation functions or in data-preprocessing part. I have two questions. Firstly, have you ever met the problem mentioned before? If yes, how to solve it. Secondly, can you tell me some detail information on your dataset such like the data distribution and the values. Or cloud you email me your datasets and the data-processing code? My email address is "[email protected]". I am quite looking forward to your reply.