Introduction
OSET is a collection of electrophysiological data and open source codes for biosignal generation, modeling, processing, and filtering. The toolbox is distributed under the GNU General Public License and may be freely used or modified under the specified terms of use. The source codes have been mainly developed in Matlab and partially in C++; but translations of the codes to other languages are welcome.

Please feel free to contact us in case of any comments or questions. You can also register in our mailing list to be informed of the toolbox modifications and its latest versions. We also welcome any contributions to this open-source project.

Packages
General filtering & processing tools
Contains a set of tools for preprocessing and analyzing biosignals such as the ECG; including R-peak detection, synchronous ensemble averaging, lowpass and bandpass filtering, baseline wander (or trend) removal, and power-line cancellation.
This package also includes nonlinear Kalman filters specifically designed for ECG/MCG filtering, which are useful for removing in-band noise from ECG/MCG recordings, or alternatively, for removing cardiac artifacts from other bioelectrical signals such as the EEG, EMG, and fetal ECG, in low SNRs.

Multichannel decomposition and filtering tools
Contains a set of tools for the decomposition and filtering of multichannel biosignals, including the ECG and MCG. The methods are based on a special class of semi-blind source separation techniques that use prior information about the desired or undesired signal and noise subspaces to decompose multichannel recordings. The prior information used by this technique can have the form of pseudo-periodicity (as for adult and fetal ECG/MCG), nonstationarity (as for EOG artifacts in the EEG), spectral contrast (as for different EEG spectral bands), event relatedness (as for ERP in EEG signals), or etc.

Multichannel synthetic biosignal and noise generator
Contains several functions for generating realistic multichannel bio-signals, including adult and fetal ECGs. This tool is also used for generating mixtures of maternal abdominal signals in single and multiple pregnancies with realistic ECG noises. The generated signals may specifically be used for studying the performance of multichannel source separation techniques such as ICA. The package also contains a set of functions for generating realistic ECG noises such as the baseline wander, muscle artifacts, and electrode movements.

Real and synthetic fetal ECG database
This package contains a set of real and synthetic fetal ECG signals recorded from maternal abdominal sensors. The real recordings have been provided by Dr. A. Tokarev from the Biomedical Signal Processing Laboratory of National Aerospace University, Kharkov, Ukraine and the synthetic ones have been produced by the multichannel synthetic ECG package of OSET. These data are specifically selected and designed for benchmarking single and multichannel signal processing algorithms.

Contributors
1- Department of Computer Science, School of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran
2- Biomedical Signal and Image Processing Laboratory (BiSIPL), Sharif University of Technology, Tehran, Iran
3- Laboratory of Images, Speech, Signals, and Control (GIPSA-lab), Institut Polytechnique de Grenoble (INPG), Grenoble, France
4- Biomedical Signal Processing Laboratory of National Aerospace University, Kharkov, Ukraine

Contact
Dr. Reza Sameni
Department of Computer Science
School of Electrical and Computer Engineering
Shiraz University, Mollasadra Street
Shiraz, Iran
Postal Code: 71348-51154
Tel: +98 (0)711 613 31 69
Fax: +98 (0)711 647 46 05
Email: info@oset.ir