الفهرس | Only 14 pages are availabe for public view |
Abstract One of the serious challenges facing health care at present is to prevent the development of epilepsy and predict epileptic seizures prematurely as earlier as possible in order to help care-givers to take appropriate precautions. Epilepsy is one of the serious diseases that affect the human nervous system. It is represented by the occurrence of recurrent, spontaneous, sudden, or unexpected glitch in activities, which leads to the occurrence of epileptic seizures. The most effective method for epileptic activity analysis among diagnostic and imaging methods is the analysis of electrical Electroencephalography (EEG) signals. Electroencephalography (EEG) is an electro-physiological technique used to track and record brain wave patterns. EEG signal processing can be utilized in various applications: medical applications such as seizure detection and prediction, and non-medical applications such as entertainment and media applications. Due to the multi-channel nature of EEG signals, channel selection is required to reduce complexity of the signal processing systems. This thesis is directed towards channel selection and seizure prediction based on statistical probability distributions of EEG signals in both time and wavelet domains. Its main idea is how to distinguish between various signal activities based on their Probability Density Functions (PDFs). Different signal attributes are investigated to anticipate the seizure onset based on the wavelet transform. These attributes include amplitude, mean, median, variance, derivative, and entropy of signals. Various wavelet families have been considered including Haar, Daubechies (db1, db4, and db8), Symlets (Sym4), and Coiflets (Coif4) wavelets. The seizure prediction process is intended to be simple to be applied on a mobile application accompanying the patient to give him alerts of possible incoming seizures. Moreover, a lossy compression technique is considered in this research based on the Discrete Cosine Transform (DCT) and Discrete Sine Transform (DST) to investigate the sensitivity of the proposed seizure prediction approach for compressed EEG signals. The first proposal achieved a sensitivity of 92.47% with false- prediction rate of 0.092/h and average prediction time of 32.52 min. for all horizons. The second proposal enhances the performance of EEG – seizure prediction system using db4 wavelet. It demonstrates better results in comparison with the first proposal. It achieves a sensitivity of 99.54% with a low FPR of 0.0818/h and a high PT of 38.1676 min. The third proposal reveals that DCT compression technique achieves better result in comparison with the DST technique achieving a sensitivity of 95.238%. These obtained results reveal that the proposed approaches can be appropriately used as a mobile application for epilepsy patients and care-givers |