For patient care, monitoring and disease diagnosis, Electrocardiogram (ECG) is one of the most important human physiological parameter which carries many embedded information about human health and especially the working and wellbeing conditions of heart. Application of ECG measurement is also very suitable for cardiac and high blood pressure patient due to its non- invasive nature. It is the graphical recording of the time varying voltages generated by the myocardium due to bioelectric activities during the cardiac cycle and representing the cyclic contraction and relaxation of the human heart muscles. Necessary information about the electrophysiology of the heart diseases and ischemic changes to the heart rhythm is provided by pure ECG signal. A cleaned ECG signal provides valuable information about the functional aspects of the heart and cardiovascular system. Diagnosis of heart diseases at an early stage can prolong human life span expectancy through appropriate treatment. Doctors find difficulties in analysing the long ECG records in short time and the human eyes are also poorly suited to detect the continuously changing morphology of ECG signal. These difficulties can be overcome by powerful computer aided diagnosis (CAD) system. The CAD system not only analyses the long ECG records and morphological changes but also provides other important features like beat detection, classification, feature extractions, arrhythmia diagnosis etc. Abnormality occurred in cardiac beats of the ECG shape is generally called arrhythmia. Arrhythmia is a common term for any cardiac disorder that differs from normal sinus rhythm. Automatic computer aided ECG signal analysis for detection of heart beat is difficult due to the large variation in morphological and temporal characteristics of ECG waveforms of different patients as well as in the same patients. The main aim of my research work is to process and extract the useful information from the ECG signal for the automatic beat detection using advance digital signal processing and pattern recognition techniques. The simple and first effective approach for cardiac beat detection from ECG signal has been the measure motivation for the work. The focus of the research is especially on increasing the detection and classification accuracy for the ECG beats and to keep the recognition performance reasonably high even in noisy conditions. The ECG beat detection and classification system consists of the following steps: pre-processing, detection of QRS complex in ECG signal, feature extraction from detected QRS complexes and classification of QRS morphologies from extracted feature set of QRS complexes using adaptive wavelet neural network to detect the cardiac arrhythmias in ECG signal.
