Analysis of Short-Time Fourier Transform (STFT) and Renyie Entropy on EEG and Speech signals

This repository contains some codes for selecting the best

• Window Length
• Overlap.

Note that window type is rectangular and FFT Window resolution is 100.

Results for Human voice signal

After applying Renyie Entropy with different window length and overlap, the result can be represented as a heatmap figure:

Then we apply best conditions (minimum Entropy) on image and show the STFT Representatin:

Results for preprocessed non-seizure EEG Signal

Results for preprocessed seizure EEG Signal

Data License

Data were collected from CHB-MIT Scalp EEG Database provided by physionet.

Open Data Commons Attribution License

Goldberger, A., Amaral, L., Glass, L., Hausdorff, J., Ivanov, P. C., Mark, R., ... & Stanley, H. E. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation [Online]. 101 (23), pp. e215–e220.