الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 148 |  |  |
| | | from | 148 | | |
Abstract
According to the world health organization and with accompanying the rise in chronic diseases, heart attack represents the first reason of death in our today world. Unfortunately, more than 80% take place in low- and middle-income countries which increases the already existed burdens. Therefore, the need of a long-term remote electrocardiogram (ECG) monitoring to detect and diagnose cardiac diseases such as irregular heartbeat (arrhythmia) is vital. Most of the existing conventional wearable systems suffers from many drawbacks, patient skin irritation, disability, hard of handling, high cost, relatively heavy weight and big size, low comfortability, and low-quality long-term signal as well. The enormous progress in textiles enables the integration of the conventional system components directly into garments. Many different designs of wearable textile ECG systems have recently emerged. A part of them is not considered as a smart garment whereas most of the others present only the electronic side of the system. Our research work introduces a comprehensive study for an improved single-lead ECG smart shirt to identify automatically the premature ventricular contraction (PVC) as a common form of arrhythmia. For artefacts-free, Marvelous Designer is implemented to design our optimized relaxed slim fit shirt. In addition, a weft knitted of 80% nylon / 20% spandex fabric is used to manufacture the inner part of the T-shirt. Moreover, light weights and small sizes electronic components are integrated to the outer part via low resistance dry stainless-steel textile electrodes and cotton fabric as an outer layer for easy transmission of weak ECG signals. For the detection of PVC, a recognition method to identify normal and PVC beats in ECG signals based on an improved template matching technique is presented. from the experimental results, it is demonstrated that the Abstract IV morphological differences in both the ventricular depolarization phase and the repolarization phase caused by PVC beats can be used to accurately detect PVCs by the variation of the correlation coefficient values of normal and PVC beats. The obtained results are very satisfactory. 12 ECG recordings from the MIT-BIH arrhythmia database are used to train our PVC classifier. 36 non overlapping recordings from the same database are also used, achieving 98.84% accuracy. In addition, 97.91%, 99.34%, 99.89%, and 98.44% accuracy are achieved with the use of INCART 12-lead arrhythmia, QT, Fantasia, and MIT BIH Supraventricular Arrhythmia databases, respectively. Another pronounced improvement for our method relies on the average maximum and minimum values of the correlation coefficient of five template beats. This enables the determination of the maximum and minimum thresholds automatically. Besides valuable accuracy improvement, our proposed approach has another vital advantage as it has low complexity and does not require expert intervention. Therefore, such an approach is to be considered more suitable for integration with smart ECG T-shirts. The heart rate measured by the T-shirt is compared to the fingertip pulse oximeter. This is due to the importance of heart rate in the detection of two other types of arrhythmias, bradycardia, and tachycardia. The evaluation of the heart rate resulted in satisfactory results compared to the fingertip pulse oximeter. Finally, the results of the analysis and the wireless communication enable the cardiologist to look after the serious cases immediately.