search

brainhack-school2020 / Biosignal-Emotions-BHS-2020

Biosignal processing for automatic emotion recognition
https://school.brainhackmtl.org/project/biosignalemotions/
Creative Commons Zero v1.0 Universal
41 stars 15 forks source link

Which features will be used #6

Closed ltetrel closed 4 years ago

ltetrel commented 4 years ago

Hi @danibene and @kabbabi,

Congrats for this clear project description, this looks promising!

I was wondering what kind of features from ECG and other biosignals you would like to extract and analyse ? For ex : the frequency spectrum, time-frequency image, wavelet transform etc..

kabbabi commented 4 years ago

Hi @ltetrel thank you for your message,

Here are the features we a priori think we will use : From ECG : RR intervals and Heart Rate Variability, Cepstral analysis, Autocorrelation matrix From EEG : Spectral analysis (Power Spectrum Density, Phase), Cepstral analysis, Autocorrelation matrix, Cross-correlation between channels, Singular spectrum analysis, Time-Frequency analysis, Wavelet Transform, Multiresolution analysis

We think we would use ECG data first to denoise EEG from cardiac artifacts, and then also to see if there are any correlations between features extracted from ECG and EEG.

Will there be any tutorials about these features ? Best, Achraf

ltetrel commented 4 years ago

Humm, I am sure nilearn has some