الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 114 |  |  |
| | | from | 114 | | |
Abstract
Electroencephalography (EEG) is a technique for detecting potentials in the brain that demonstrate electrical activity. It is a simple test that demonstrates how the brain functions over time. EEG recordings are one of the most useful tools for diagnosing neurological diseases, including epilepsy, brain tumors, autism and sleep disorders. EEG has been a highly effective laboratory investigative technique for identifying seizures and epilepsies for decades. The importance of the EEG is in its ability to carry out multiple diagnosis tasks such as the presence of epilepsy, the location of potential epileptic foci, the identification of the epileptic zone for surgical resection in intractable epilepsies, the prognosis of epileptic surgery, the impacts of medication, an assessment of drug safety and toxicity. This dissertation presents two different techniques for online epilepsy diagnosis. The first technique presents a new automated on-line method for EEG epileptic seizure diagnosis based on multi- stage of Quantized Kernel Least Mean Square (QKLMS) adaptive filters. The QKLMS filter is utilized in the suggested framework to achieve accurate real-time detection with low hardware complexity. In the suggested multi- stage filter design, the energy of predicted signal is investigated to detect the seizure interval on the EEG records and to classify healthy, epileptic interictal, and epileptic ictal cases. The Grey Wolf Optimizer (GWO) algorithm is used to find the optimum values of the energy threshold and the parameters of the QKLMS algorithm The performance of the proposed QKLMS technique is compared with other adaptive filters, and the results reveal the superior performance of the QKLMS algorithm. The proposed framework is examined with real EEG records taken from the Bonn University database, and the experimental results show that the proposed epileptic seizure diagnosis approach outperforms other state-of-the-art systems with an accuracy of 97.88 %, sensitivity of 98.80 %, specificity of 97.65 %, and computational time of 0.58 sec.
The second technique proposes an automated Deep Learning (DL) approach based on integrating a pre-trained Convolutional Neural Network (CNN) structure, called AlexNet, with the Constant-Q Non-Stationary Gabor Transform (CQ-NSGT) algorithm for classifying seizure versus seizure-free EEG records. The CQ-NSGT method is introduced to transform the input 1-D EEG signal into 2-D spectrogram which is sent to the AlexNet CNN model. The AlexNet architecture is used to obtain the discriminating features of the 2-D image corresponding to each EEG record in order to distinguish seizure and non-seizure cases using Multi-Layer Perceptron (MLP) algorithm. The robustness of the proposed CQ-NSGT technique in transforming the 1-D EEG signals into 2-Dkpectro grams is assessed by comparing its classification results with the Continuous Wavelet Transform (CWT) method, and the results prove the high performance of the CQ-NSGT technique. The suggested system is investigated with the same data used in the first approach, and the experimental results reveal the superior performance of the proposed framework over other state-of-the-art methods with an accuracy of 99.56 %, sensitivity of 99.12 %, specificity of 99.67 %, and precision of 98.69 %.