Volume 5, No. 8, Nov-Dec 2014 International Journal of Advanced Research in Computer Science RESEARCH PAPER Available Online at www.ijarcs.info © 2010-14, IJARCS All Rights Reserved 114 ISSN No. 0976-5697 Two-stage Nonlocal Means Denoising of ECG Signals Ambuj Dubey PG Student,Deptt.of Electronics& communication Engg, Oriental Institute of Science and Technology Bhopal(M.P.) Prof. Hasnine Mirza Deptt. Of Electronics & Communication Engg; Oriental Institute of Science & Technology Bhopal(M.P.) Prof. M.Ahmed HOD Deptt. Of Electronics & Communication Engg; Oriental Institute of Science & Technology Bhopal(M.P.) Abstract: Every type of biomedical signals is generated by physical activities in the body. In a present scenario some attention has been created by medical biometrics, including ECG (Electrocardiogram), DNA, blood pressure, EEG (Electroencephalogram) and heart rate [1]. An electrocardiogram (ECG) uses to trace and describes heart electrical activity and it recorded by electrodes placed on the body surface. In this paper, we study various filters which have been implemented for reduction of noise in ECG. Their performances are also compared based on the SNR values. But problem still same that noise can overlap the entire signal, so these cases the classical methods in signal denoising are not acceptable. So reduce that difficulty we propose a noval approach which based on the Non Local Mean (NLM) algorithm. The NLM was recently introduced in as a technique for processing nonlinear and non stationary signals. Keywords: ECG signal, non-local means, wavelet and Denoising. I. INTRODUCTION ECG signal is a combination of PQRST waves. ECG used for detect heart related diseases where P wave , QRS wave and T wave are deferent function The electrocardiogram (ECG) is the recording of the cardiac activity and it is extensively used for diagnosis of heart diseases. It is also an essential tool to allow monitoring patients at home, thereby advancing telemedical applications. Recent contributions in this topic are reported in [2–4]. Inside the clinical environment under observation of ECG signals, the ECG signal having various types of noise or artifacts. The spread of ECG signal introduces noise or artifacts because of the poor channel conditions. In ECG enhancement, the goal is to separate the valid ECG from the undesired artifacts so as to present a signal that allows easy visual interpretation. Many approaches have been reported in the literature to address ECG enhancement. Some recent relevant contributions have proposed solutions using a wide range of different techniques. In this letter, we briefly describe the NLM algorithm in second stage, and discuss its application in the context of ECG denoising. Here we present results for ECG denoising of signals with simulated additive noise. Positively, the NLM algorithm results are very best with compare to recently published wavelet denoising[5]. II. NOISES IN ECG SIGNAL ECG measurements may be corrupted by many sorts of noise. The ones of primary interest are[3]: a. Power line interference b. Electrode contact noise c. Motion artifacts d. EMG noise e. Instrumentation noise f. AWGN These artifacts strongly affects the ST segment, degrades the signal value, frequency resolution, produces great amplitude signals in ECG that can resemble PQRST waveforms show in fig.1, and masks tiny features that are important for clinical monitoring and diagnosis. Cancelation of these noises in ECG signals is an important task for better diagnosis. Many sources of signal contamination including additive high frequency noise (AWGN), motion or muscle artifacts, and baseline wander overlap signals of clinical interest in both time and frequency. Figure.1. ECG signal III. ECG DENOISING ALGORITHMS Noise cancellation in ECG signal requires different- different strategies for different noise sources or types. However, in real situations ECG signals recordings are mainly highly corrupted by artifacts. Here basically two leading artifacts existing in ECG recordings which are: (1) high-frequency noise which produced by electromyogram induced noise, power line interferences, or mechanical forces acting on the electrodes; (2) baseline wander disturbances (BW). There are many useful methods for removing power line and baseline wander disturbances in our ECG signal by digital linear phase filtering [6]. This method can be used to reduce signal magnitude spectrum while preserving the signal time domain as much as possible. The disadvantage of this method is the computational requirements. This is mainly caused by a