الفهرس 
TitleOnly 14 pages are availabe for public view
 |  | from | 86 |  |  |
| | | from | 86 | | |
Abstract
Summary
5.1.1. Introduction and aim of the work
The heart disease is one of the live threaten diseases and top urgent to any cardiologist. Diagnosing the heart cases accurately and readily will save a lot of patients. Automatically diagnosis those cases will give the cardiologists a remarkable help to can do their jobs easily. So the heart examination devices are considered the more important innovation in the last centuries. The electrocardiogram (ECG) device is one of those examination devices and the most common device between all the heart cases as a tool of a diagnosis helping the cardiologist to take his decision.
The ECG device takes its reading from the patient by leads attached on the human body. This reading is a called the ECG signal. It is an electrical signal sensed by the leads from the human body by electrical sensors in those leads. This signal is processed, showed and stored by the ECG device. By this ECG signal the cardiologist can create his opinion about the case. The more updated the ECG device and developed is, the more accurate decision the cardiologist can take. So this novel hybrid system are focused in the ECG signal development to help more the cardiologist.
5.1.2. Review of Literature
The Electrocardiogram is one of the important devices that researchers focus their work on. A lot of researches had been made to improve the quality and the accuracy of the ECG signals. Some works focused on the filtration or denoising phase to get a more pure ECG signals. Others targets the detection and diagnosing phase by using different techniques from traditional scaling, transforming or neural network techniques. Few researches mixed between wavelet and neural network techniques in the processing phase. This mixture was very complicated and gave less accurate results with long time processing. None of the researchers mixed between different techniques in the preprocessing phase. Those previous works were still not get the cardiologists’ confidence to base their diagnoses on. So this proposed work aimed to get a better results to get the cardiologists’ trust to save more time in diagnosing and patient lives.
55
5.1.3. Materials and methods
In this system a mixture between two different techniques in the preprocessing is implemented more pure ECG signal. The first used technique is the filtration technique which removes the external noises around the ECG signal. The external noises are the power line noise, baseline and the DC shift. The second used technique is the denoising which selects only the ECG signal frequency band and eliminates the other internal noises. The internal noises is found by the lung and the heart muscles movements. This system mixture between those two techniques as an advantage of this system to give a pure signal for the second step of detection and diagnosis.
The detection phase is developed too by adding a novel scaling technique. This technique is called time scaling technique which helps more in detecting the cases of the ECG signal not only by its value but also but its timing and location on the ECG signal timeline. This novel technique gives the system the ability to diagnosis more cases that were so hard to be diagnosed before. Those cases are the Premature and Bigeminy cases, and diagnosed the cases which previously detected in the last works with higher accuracy.
This system also can overcome one of the most common errors in the ECG devices which is the double counting. The ECG devices can double count the heartbeats because of the split R-wave, small R or the large positive T wave. This novel hybrid system and with the novel time scaling technique can cancelling this errors and diagnose the ECG signal on a right base.
The Flat or Negative T is detected and diagnosed by this system as it wasn’t detected or diagnosed before. In the traditional scaling works, it was detected and diagnosed as a Q or S waves as it didn’t depend on the frequency and depended only the amplitude scaling for this wave. In the wavelet technique works, it didn’t detected or diagnosed because it was considered as noises.
The five new cases (premature, Bigeminy, double counting, small R and Flat or Negative T) are detected and diagnosed easily and with high accuracy results. This is considered a new effective addition to the automated ECG signals diagnosis.
This novel hybrid system does not based only on the new time scaling technique. It based also on the traditional scaling amplitude value technique and the Wavelet decomposition technique. This mixture what make this system results a higher accuracy for the cases than previous works and detect more new cases.
This mixtures between different techniques in the preprocessing phase and the processing phase does not affect the processing time. Like, it is obvious the consumption time is very small comparing to that much of processing on the ECG signal to get this high accurate result. This is another advantage of this novel hybrid system.
Another final advantage of this system, that it is a self-calculated system for the accuracy. According to the cardiologist opinion a database for the accuracy is created and compared with the system results. An automatic calculation for the result accuracy is displayed after the diagnosis part to give the cardiologist the clear picture for this results.
56
5.1.4. Results and discussion
The Results of this novel hybrid system are promising. The ECG signals that get from this system is a very pure signal because of the mixing between 2 different techniques for the first time. The accuracy of the detection and diagnosis phase has high percentage with an average equals 97.7% with less possible processing time equals in average 0.05 sec. A new 5 cases are detected and diagnosed easily with the help of this system in contrast with the previous works. The cases that were detected and diagnosed before by the previous works, are detected and diagnosed by this novel hybrid system with higher accuracy.
5.2. Conclusion
ECG signal filtration, denoising, detection, feature extraction and diagnosis are one of the important topics nowadays. Biomedical engineering field is full of research around that point in order to get a more accurate result for interpreting cardiac activity. Heart diseases can be in the heart rate, amplitude or in the rhythm of ECG waves. In this work a hybrid system between a modified scaling and Wavelet transform techniques is proposed in which the amplitude and the time scaling technique are combined for computer interpretations ECG with high accuracy and less running time. Utilizing the signal time in modified technique allows to detect many episodes unrealized by the traditional scaling such as regularity (premature) and bigeminy. Five new cases are detected and diagnosed by this novel hybrid system. Furthermore, it solved the traditional errors such as double counting and split R-wave errors. For this technique, the average accuracy percentage was 97.7 % and the average time consumption was 0.0503004 seconds. After presenting the above results, this system is considered a trusted technique to cardiologists which they can based their decisions on.
The proposed method improved the performance of the filtration and denoising phase. The mixture between different techniques (the traditional scaling with the modification of the time scaling technique and wavelet transform technique) helps in this improvement in results. The modification in scaling technique by adding the time scaling technique is one of the important factor in the detection phase. Also, it improved the performance and the accuracy of the detection and diagnosis phase. It can be used as a second opinion with the cardiologists to take the right decision for the patient’s case with in the right time.
5.3. Recommendations
All the above advantages make this novel hybrid system is so promising and helpful to the cardiologists and to diagnose the heart cases to save lives. Also this system can be used as an educational tool for the new cardiologists to understand the electrical activity of the heart, practice more and improve their skills. This system can help also on the ambulances in the absence of the cardiologist because of its higher accuracy results. In addition, it can be connected with the hospital network to update the cardiologist with the upcoming case to save time and start diagnosing the case before even the arrival of the patient. Those suggestions are considered as the future work that can be built on this novel hybrid system to develop the ECG device more and more.