138 | Page ECG BEAT CLASSIFICATION USING CROSS- WAVELET AND LVQ Priyadarshiny Dhar 1 , Dr. Saibal Dutta 2 , Dr.Prithwiraj Das 3 , Dr. Vivekananda Mukherjee 4 , Abhijit Dhar 5 1 Applied Electronics & Instrumentation Engineering, Netaji Subhash Engineering College, Panchpota,Garia, Kolkata-700152, West Bengal, (India) 2 Electrical Engineering, Heritage Institute of Technology,Chowbaga Road, Anandapur, PO:East Kolkata Township, Kolkata 700 107, West Bengal, (India) 3 Associate Professor, Govt. College of Engineering and Textile Technology, Berhampore, 742101,West Bengal, (India) 4 Department of Electrical Engineering, Indian School of Mines, Dhanbad - 826 004, Jharkhand, (India) 5 Consultant, BI & Data Analytics Architect, Tata Consultancy Services Ltd. West Bengal, (India) ABSTRACT This paper describes an automatic classification system based on combination of cross-wavelet and Learning Vector Quantization (LVQ) for the purpose of automatic heartbeat detection. The feature extractor is based on cross- wavelet approach, using the time frequency information. The ANN classifier uses a Learning Vector Quantization (LVQ) method which classifies the ECG beats into two categories: normal beats and abnormal beats. The ECG (electrocardiogram) signals in the MIT-BIH arrhythmia database are adopted as reference data. Total 98530 heart beats are used for testing the above classifier. The total classification accuracy (TCA) was about 91.66%. Keywords: Cross-Wavelet Spectrum, Cross-Wavelet Coherence Spectrum , LVQ I INTRODUCTION The electrocardiogram (ECG) is the recording of the electrical property of the heartbeats, and has become one of the most important tools in the diagnosis of heart diseases. Due to the high death rate of heart diseases, early detection and precise discrimination of ECG arrhythmia is essential for the treatment of patients. This requisite contributes to intensive studies in recent years for high-precision computer-aided diagnosis (CAD) systems for ECG. An effective CAD system requires a powerful pattern classifier as well as a graceful feature extractor that is capable of extracting