Proficient Technique for Classification of ECG Signal

ECG's are one of the most important biomedical signals; these are usually used to determine the insightful of an electric activity of the heart. ECG gives the statistics about the electrical functionality of heart, with help of its constituent waves shape, that is the P, QRS, and T waves. Features of the ECG have a noteworthy role in detection of the various cardiac syndromes. PQRS- T waves are the components of one cycle of ECG signal. The heart functionality can be determined with the help of the intervals and amplitude of the P -QRS-T segment. This paper proposes an approach to examine electrocardiogram (ECG) signal. The working of proposed classifier is based artificial neural network (ANN), In this feature extraction is done by means of discrete wavelet transform (DWT) whereas for classification purpose neural network is used and the classification is done in five different types of arrhythmias viz. Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), Paced Beat (PB), Atrial Premature Beat (APB) and First degree AV Block (AVB) beats apart from normal (NS) beats. The anticipated neural network (NN) based comprehensive classifier gives an enriched performance of classification the performance measures obtained are: sensitivity is95%, specificity is 99.01% and classification accuracy is98.35%. Keywords: ANN (Artificial Neural Network), ECG (Electrocardiogram), discrete wavelet transform (DWT)