الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
The human heart is the one of most important topic in medical researches. In recent years there arc many diagnose techniques of the heart in which the ECG plays the important role. Among these techniques the artificial intelligent occupies an increasing attention. Therefore, the objective of this thesis is to introduce proposals that facilitate the study and analysis of the ECG complex wave signal using the neural network to derive heart diseases.
The study is based on three approaches. A proposal for preprocessing the ECG complex wave signal had been introduced. The preprocessing enables to remove noise and any artifact from the hard copy signal, moreover, it helps to reconstruct a noise free original ECG signal. Finally, parametric correlation ship is applied to extract the true QRS wave that is used to conclude patient disease.
The determination of cardiac axis has been calculated using three methods (bipolar, unipolar, and hybrid leads configurations), From the stLldy of the obtained results it was shown that the introduced proposal of the hybrid leads method is the simplest method.
A second, whole vector diagram, proposal has been introduced that takes in consideration the whole ECG complex wave instead of considering the QRS wave only. This proposal shows a slight change in the determination of the CA deviation with respect to the other methods.
Finally, an integrated ECG 12-lead diagnosis is introduced. In this proposal the 12-lead ECG signals are initially preprocessed, time division multiplexed, data compressed and analyzed by pretrained artificial neural network (ANN). The introduced ANN can reject any disease that is not trained before or recognizes a new disease and adds it to teacher. This proposal allows to use ANN for real time diseases diagnosis.