چکیده:
Electrocardiogram (ECG) is a tool used for the electrical analysis of the status of human heart activity. When the ECG signal is recorded, it gets contaminated with different types of noises. So, for accurate analysis, noises must be eliminated from the ECG signal. There are different types of noises that contaminate the characteristics of ECG signal i.e Power line interference, baseline wander, Electromyogram (EMG). In this paper, different techniques have implemented for the removal of noises. A median filter is used for removal of DC component and Savitzky-Golay filter (SG) is used for smoothing noised waveform and then wavelet transform (db4) is used to decompose the ECG signal for removal of various artifacts. Wavelet transform provides the information in frequency and time domain and then thresholding has been applied for the implementation of algorithms in MATLAB. The measured results i.e. SNR(Signal to Noise ratio) and MSE(Mean square error) have been calculated using different databases like MIT-BIH, Long-term ST database, European ST-T database. The results are examined with proposed methods that are better than those reported in the literature.
خلاصه ماشینی:
A median filter is used for removal of DC component and Savitzky-Golay filter (SG) is used for smoothing noised waveform and then wavelet transform (db4) is used to decompose the ECG signal for removal of various artifacts.
Base line wander, ECG, EMG, MSE (Mean square error), Power line interference, Savitzky-Golay filter, Signal to Noise Ratio (SNR), Wavelet transform / DOI: 10022059/jitm.
In (Eminaga, 2018), a wavelet transforms and hybrid IIR/FIR filter is proposed for the removal of noises from the ECG signal.
After that wavelet transforms(db4) is used for elimination of high frequency noises present in electrocardiogram (ECG) signal.
The proposed method is validated on standard database of European ST-T database, Long-term ST database and MIT-BIH database for different records and measured results in the form of Mean square error (MSE) and Signal to noise ratio (SNR), and compared these results with existing works.
After that output of median filter is passed to Savitzky-Golay (SG) filter for smoothing of noise waveform and then wavelet transform (db4) has been employed to decompose the electrocardiogram (ECG) signal and eliminate the high frequency noises.
Results showing removal of baseline wander for different records of different database a) 100 b) 101 c) e0103 d) e0104 e) S20011 f) S20021 The range of frequency of this artifact is 005 to1 Hz. In this proposed work, the electrocardiogram (ECG) signal is decomposed up to tenth level into approximation co- efficient and detail co-efficient (Manocha and Singh, 2016).