nnsa.feature_extraction.brainagemodel.models package
Submodules
nnsa.feature_extraction.brainagemodel.models.sincnetwork module
The Sinc Keras model presented in the paper. To get the model call ‘net’ function.
Functions:
|
This function returns the Sinc network presented in the paper. |
- nnsa.feature_extraction.brainagemodel.models.sincnetwork.net(config)[source]
This function returns the Sinc network presented in the paper.
Parameters:
- config: an object from any class (preferably from brainagemodel.core.config.Config) which includes the following attributes:
fs: sampling frequecy in Hz (according to the paper, it should be 64Hz)
frame_sec: defining the frame length of EEG in seconds. According to the paper it should be 30s.
CH: the number of EEG channels. According to the paper it should be 1, 2, 4, or 8.
- Example:
model = net(Config(fs=64, CH=8, frame_sec=30))
The input signal to the network will be [frame_sec*fs), ch] (paper: [1920(30s*64Hz), 8/4/2/1])
Return:
a Keras API-based model (not compiled).