1887

Abstract

Abstract

Electroencephalogram (EEG) is a representative signal containing information of the electrical activity generated by the cerebral cortex nerve cells; it has been the most utilized signal to clinically assess brain activities, and to detect abnormalities such as epilepsy. However, the manual detection of such brain abnormalities as epilepsy or seizure includes visual scanning of EEG recordings, which is very time consuming especially in the case of long recordings. So, the EEG signal parameters are extracted and analyzed using computer based digital signal processing techniques are highly useful in diagnostics and more suitable for detecting and classifying EEG abnormalities.

This work aims to develop novel features extracted from the time-frequency distribution (TFD) of the EEG signals including newborn for the purpose of classifying these signals in three possible categories i.e.: 1) acquired from healthy subjects, or 2) epileptic patients during normal brain activity, or 3) epileptic patients while experiencing seizures.

The proposed method for classifying EEG signals includes the three following stages:

- Time-frequency decomposition of EEG signal using the quadratic time-frequency representations (TFR).

- Features extraction from TFR.

- Features classification in order to assign the signal to one abnormality class: mild, moderate or severe abnormality.

The experimental results show that the proposed method provide better results using certain types of Quadratic TFDs such as the Modified-B distribution or the Spectrogram distribution in combination with the support machine classifier to detect and classify the epilepsy. We also found that the performance of this method is not related only to the TFR and the classifier choice but is also dependent on the choice of significant features. We are currently developing new image processing techniques to extract new features from the TFR considered as an image. The design is based on the use of edge and contour detection, and, segmentation methods in order to define new features such as the number, the shape and the localization of the components.

Loading

Article metrics loading...

/content/papers/10.5339/qfarf.2011.BMP12
2011-11-20
2020-09-27
Loading full text...

Full text loading...

http://instance.metastore.ingenta.com/content/papers/10.5339/qfarf.2011.BMP12
Loading
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error