In this paper an algorithm has been proposed to detect and classify the cardiac arrhythmia from a normal Electro Cardio Graphic (ECG) signal based on wavelet decomposition with adaptive threshold. The MIT -BIH arrhythmia and malignant ventricular arrhythmia database has been utilized for evaluating the algorithm. The performance of the algorithm is compared with some existing algorithms in terms of signal duration time (episode length), sensitivity, specificity and positive selectivity. The analysis shows that the proposed algorithm gives satisfactory results.

No abstract