Abstract
We propose a new method for detection and classification of noisy recorded epileptic transients in Electroencephalograms (EEG) using the continuous wavelet transform (CWT) and artificial neural networks (ANN). The proposed method consists of a segmentation, feature extraction and classification stage. For the feature extraction stage, we use best basis mother wavelet functions and wavelet thresholding technique. For the classification stage, multilayer perceptron neural networks were implemented according to standard backpropagation learning formulations. We demonstrate the efficiency of our feature extraction method on data to improve the ANN detection performance. As a result, we achieved the accuracy in detection and classification of seizure EEG signals with 94.69%, which is relatively good comparing with the available algorithms at present time.
Original language | English |
---|---|
Pages (from-to) | 33-48 |
Number of pages | 16 |
Journal | International Journal of Wavelets, Multiresolution and Information Processing |
Volume | 8 |
Issue number | 1 |
DOIs | |
Publication status | Published - Jan 1 2010 |
Keywords
- Artificial neural networks
- Continuous wavelet transforms
- Epileptic seizure detection
ASJC Scopus subject areas
- Signal Processing
- Information Systems
- Applied Mathematics