Classification of ECG arrhythmias using multi-resolution analysis and neural networks

Automatic detection & classification of cardiac arrhythmias is important for diagnosis of cardiac abnormalities. We propose a method to accurately classify ECG arrhythmias through a combination of wavelets and artificial neural networks (ANN). The ability of the wavelet transform to decompose signal at various resolutions allows accurate extraction/detection of features from non-stationary signals like ECG. A set of discrete wavelet transform (DWT) coefficients, which contain the maximum information about the arrhythmia, is selected from the wavelet decomposition. These coefficients in addition to the information about RR interval (the difference between the present and previous QRS peaks) are fed to the back-propagation neural network which classifies the arrhythmias. In the present work the ECG data is taken from standard MIT-BIH Arrhythmia database. The proposed method is capable of distinguishing the normal sinus rhythm and 12 different arrhythmias. The overall accuracy of classification of the proposed approach is 96.77%. The results of the analysis were found to be more accurate than those of the existing methods. To check the robustness of the algorithm, three types of noise, i.e., muscle noise, power-line interference and base-line wander, were added with SNR values ranging from 0 dB to 10 dB to the signal and the accuracy was found to be well within the clinical limits. It was observed that the effect of base-line wander on the accuracy of detection was less than the other disturbances.