الفهرس Only 14 pages are availabe for public view
 |  | from | 112 |  |  |
| | | from | 112 | | |
Abstract
This thesis presents an overview about the recent electrocardiogram (ECG) automated diagnoses techniques. It was found that most techniques are depending on the artificial neural network (ANN) to diagnose different ECG diseases, using ECG diagrams only, and the other using phonocardiograph (pCG) only. The present thesis introduces an integrated diagnosis of the heart diseases, using an introduced integrated cardiograph scanned images (ICGI), for the ECG, and the PCG scanned images combination using ANN. The introduced method is composed of a hybrid two analyzing techniques: statistical and decompositional. This introduced method is to study and analyze the ECG scanned images, the PCG scanned images, and finally the lntegrated ICG scanned images. The extracted features (pre-extracted ~eatures, energy features, post-extracted features) are used as an input to the ntroduced ANN classifier. The architecture of the introduced ANN is lptimized based on standard measures. Finally the introduced ANN classifier ~as tested for all input features to obtain the higher classification accuracy.
: is found that: In case of using ECG scanned images only for diagnosis, the ~e-extracted features based on gray level histogram method, and the energy atures based on wavelet decomposition method leads to an accuracy of t.41 %, where as in case of using PCG scanned images only for diagnosis, ~ pre extracted features based on fast F ourier transform method & the post tracted features based on edge detection method, leads to an accuracy of .82%.
case of integrated diagnosis Of ICG scanned images, the pre extracted tures using gray level histogram & discreet cosine transform methods leads m accuracy of96.82 %.
The present thesis consists of five chapters; that can be summarized as in the following:
Chapter 1:
It presents a summary for the recent ECG and PCG automate diagnosis.
Chapter 2:
This chapter deals with the ECG diagnoses in cardiology, and th diseases that can be diagnosed from it.
Chapter 3:
This chapter deals with the heart sounds diagnoses (PCG) i cardiology. It shows that, heart diagnosis is not only concerned with studyin ECG, but also it needs to be diagnosed by the PCG.
Chapter 4:
This chapter introduces an integrated diagnosis of the heart disease using integrated cardiograph scanned images (ICGI), based on ANN. Man methods were used to extract the features from the used scanned images 0 the different heart diseases. The introduced methods are divided statistical methods (edge detection, and histogram methods), decompositional methods (wavelet decomposition, fast Fourier, and discree cosine transforms). The extracted features are ”pre extracted features, energ features, and post extracted features” used as an input to the introduced AN: classifier. The introduced ANN was optimized based on standard measures. Finally the introduced ANN classifier was tested for all input features t obtain the higher classification accuracy.
Chapter 5:
It summaries the main conclusions that obtained in this thesis an introduces some future works.