I am working on an EEG signal classification problem. My dataset consists of EEG signals stored as 19X30000 NumPy arrays. Each row represents a single channel. For now, I am converting each of the individual rows into a spectrogram and concatenating all these 19 spectrograms to produce a single image representing the EEG recording that will be fed to a CNN later.
I wanted to know if there is any other way of doing it. Is there any way by which I can represent the 19X30000 input array with a 1X30000 array and use spectrogram of that? I once read about convoluting first across time and then across the channel to reduce it to a single array but I cant remember it exactly. Can someone help me with this?