An adaptive level dependent wavelet thresholding for ECG denoising

On the basis of wavelet theory, a novel Adaptive wavelet thresholding method (AWT) is proposed for the ECG signal enhancement. The best base wavelet for ECG signal filtering can be automatically obtained through the cross correlation coefficient and the energy to entropy ratio. The variable universal threshold (VarUniversal) is applied to different decomposition level so as to suppress diverse noise. To achieve a smooth cut-off transition, an identical correlation shrinkage function (IcoShrinkage) is also adopted in the AWT according to its correlation coefficients with the hard thresholding and the soft thresholding. The performance of AWT is compared with four threshold approaches and six shrinkage functions, respectively, on the basis of 150 practical ECG signals of 30 subjects. The filtering results reveal that the AWT can adaptively choose an optimal base wavelet for a specific ECG signal. With the VarUniversal threshold and IcoShrinkage, the AWT obtains the better filtering results than the other compared methods.

Robust, real-time event detection from physiological signals acquired during long-term ambulatory monitoring still represents a major challenge for highly-artifacted signals. In this paper, we propose an original and generic multi-feature probabilistic detector (MFPD) and apply it to real-time QRS complex detection under noisy conditions. The MFPD method calculates a binary Bayesian probability for each derived feature and makes a centralized fusion, using the Kullback-Leibler divergence. The method is evaluated on two ECG databases: 1) the MIT-BIH arrhythmia database from Physionet containing clean ECG signals, 2) a benchmark noisy database created by adding noise recordings of the MIT-BIH noise stress test database, also from Physionet, to the MIT-BIH arrhythmia database. Results are compared with a well-known wavelet-based detector, and two recently published detectors: one based on spatiotemporal characteristic of the QRS complex and the second, as the MFDP, based on feature calculations from the University of New South Wales detector (UNSW). For both benchmark Physionet databases, the proposed MFPD method achieves the lowest standard deviation in sensitivity and positive predictivity (+P) despite its online algorithm architecture. While the statistics are comparable for low-to mildly artifactual ECG signals, the MFPD outperforms reference methods for artifacted ECG with low SNR levels reaching 87.48 ± 14.21% in sensitivity and 89.39 ± 14.67% in +P as compared to 88.30 ± 17.66% and 86.06 ± 19.67% respectively from UNSW, the best performing reference method. With demonstrations on the extensively studied QRS detection problem, we consider that the proposed generic structure of the multi-feature probabilistic detector should offer promising perspectives for long-term monitoring applications for highly-artifacted signals.

Physiological variation of the interval between consecutive heartbeats is known as the heart rate variability (HRV). HRV analysis is traditionally performed on electrocardiograms (ECG signals) and has become a useful tool in the diagnosis of different clinical and functional conditions. The progress in the sensor technique encouraged the development of alternative methods of analyzing cardiac activity: Seismocardiography and gyrocardiography. In our study we performed HRV analysis on ECG, seismocardiograms (SCG signals) and gyrocardiograms (GCG signals) using the PhysioNet Cardiovascular Toolbox. The heartbeats in ECG were detected using the Pan–Tompkins algorithm and the heartbeats in SCG and GCG signals were detected as peaks within 100 ms from the occurrence of the ECG R waves. The results of time domain, frequency domain and nonlinear HRV analysis on ECG, SCG and GCG signals are similar and this phenomenon is confirmed by very strong linear correlation of HRV indices. The differences between HRV indices obtained on ECG and SCG and on ECG and GCG were statistically insignificant and encourage using SCG or GCG for HRV estimation. Our results of HRV analysis confirm stronger correlation of HRV indices computed on ECG and GCG signals than on ECG and SCG signals because of greater tolerance to inter-subject variability and disturbances.

Automated ECG signal processing can assist in diagnosing several heart diseases. Many R peak detection methods have been studied because the accuracy of R peak detection significantly affects the quality of subsequent ECG feature extraction. Two important steps in R peak detection algorithm that draw attention over researchers are the preprocessing and thresholding stages. Among several methods, wavelet transform is a widely used method for removing noise in the preprocessing stage. Various proposed algorithms require prior knowledge of frequency spectrum of the signal under consideration in order to select the wavelet detail coefficients in the reconstruction process. Moreover, parameter fine tuning is generally involved in threshold selection to accomplish high detection accuracy. As a result, it may be difficult to utilize these methods for general ECG data sets. Accordingly, we propose an automatic and parameter free method that optimally selects the appropriate detail components for wavelet reconstruction as well as the adaptive threshold. The proposed algorithm employs the analysis of probability density function of the processed ECG signal. The validation of the algorithm was performed over the MIT-BIH database and has produced an average sensitivity of 99.63% and specificity of 99.78% which is in the same range as the previously proposed approaches.

In this paper, an efficient Fast Block LMS (FBLMS) algorithm is proposed for removing artifacts preserving the low frequency components and tiny features of the ECG. The proposed implementation is suitable for applications requiring large signal to noise ratios with fast convergence rate. The FBLMS algorithm, being the solution of the steepest descent strategy for minimizing the mean squared error in a complete signal occurrence, is shown to be steady-state unbiased and with a lower variance than the LMS algorithm. Finally, we have applied this algorithm on ECG signals from the MIT-BIH data base and compared its performance with the conventional LMS algorithm. The results show that the performance of the FBLMS algorithm is superior than the LMS algorithm. I. INTRODUCTION The extraction of high-resolution ECG signals from record- ings contaminated with background noise is an important issue to investigate. The goal for ECG signal enhancement is to separate the valid signal components from the undesired artifacts, so as to present an ECG that facilitates easy and accurate interpretation. Many approaches have been reported in the literature to address ECG enhancement using adaptive filters (2)-(3), which permit to detect time varying potentials and to track the dynamic variations of the signals. In (3), Thakor et al. proposed an LMS based adaptive recurrent filter to acquire the impulse response of normal QRS complexes, and then applied it for arrhythmia detection in ambulatory ECG recordings. The reference inputs to the LMS algorithm are deterministic functions and are defined by a periodically extended, truncated set of orthonormal basis functions. In these papers, the LMS algorithm operates on an instantaneous basis such that the weight vector is updated every new sample within the occurrence, based on an instantaneous gradient estimate. There are certain clinical applications of ECG signal processing that require adaptive filters with large number of taps. In such applications the conventional LMS algorithm is computationally expensive to implement. The block processing of data samples can significantly reduce the computational complexity. By applying this strategy, a special implementa- tion of the LMS algorithm is called the block LMS algorithm. In a recent study, however, a steady state convergence analysis for the LMS algorithm with deterministic reference inputs showed that the steady-state weight vector is biased, and thus, the adaptive estimate does not approach the Wiener solution. To handle this drawback another strategy was considered for estimating the coefficients of the linear expansion, ie., the BLMS algorithm (8), in which the coefficient vector is updated only once every occurrence based on a block gradient estimation. The BLMS algorithm has been proposed in the case of random reference inputs and has, when the input is stationary, the same steady state misadjustment and convergence speed as the LMS algorithm. A major advantage of the block, or the transform domain, LMS algorithm is that the input signals are approximately uncorrelated. Moreover, the filter output and the weight update terms can be evaluated faster using FFT-based fast BLMS (FBLMS) algorithm. The advantage of this algorithm is less computational complexity and good filtering capability. These characteristics may plays a vital role in biotelemetry, where extraction of noise free ECG signal for efficient diagnosis and fast computations, high data transfer rate are needed to avoid overlapping of pulses and to resolve ambiguities. To the best of our knowledge, transform domain has not been considered previously within the context of filtering artifacts in ECG signals. In this paper, we present a FBLMS algorithm to remove the artifacts from ECG. Such a realization is intrinsically less complex than its BLMS based counterpart. This algorithm enjoys less compu- tational complexity and good filtering capability. To study the performance of the proposed algorithm to effectively remove the noise from the ECG signal, we carried out simulations on MIT-BIH database for different artifacts. The simulation results shows that the proposed algorithm performs better than the LMS counterpart to eliminate the noise from ECG.

This paper presents a state-of-the-art review of the recent advances in computer methods for analysis and design of switched circuits in both time and frequency domains. It also presents new research results on time-domain analysis and sensitivity of switched nonlinear circuits, time- and frequency-domain statistical analysis of switched linear and nonlinear circuits, and efficient modeling and analysis of clock feedthrough and charge injection in switched linear circuits. In time-domain analysis, the modeling of switches and its effect on the simulation of switched circuits are investigated. Formulation methods for these circuits are examined. Inconsistent initial conditions arising from ideal switching are investigated and numerical methods that derive the consistent initial conditions are examined. Sampled-data simulation (SDSIM) of switched linear circuits including clocked sigma-delta modulators is investigated. SDSIM is extended to switched nonlinear circuits. Time-domain sensitivity of switched linear and nonlinear circuits is analyzed using SDSIM. Efficient time-domain statistical analysis of switched linear and nonlinear circuits is introduced, and their effectiveness is assessed using Monte Carlo simulation. Methods that compute the effect of the clock jitter of periodically switched linear (PSL) circuits are examined. Time-domain noise analysis of PSL circuits is investigated. In frequency-domain analysis, exact frequency analysis of multiphase PSL circuits is reviewed. Sensitivity analysis of these circuits is examined in detail. Adjoint network theory and its usefulness in noise and sensitivity analysis of switched linear circuits are studied. Group delay of PSL circuits is investigated briefly. Efficient modeling and analysis of clock feed-though and charge injection of PSL circuits are introduced. Distortion and sensitivity of periodically switched nonlinear circuits with mild nonlinearities are investigated. Finally, frequency-domain noise analysis of PSL circuits is examined.

The denoising of medical images degraded by various types of noises is a major problem in medical image processing. A very good tradeoff between detail preservation and noise removal is obtained while implementing with various successive filtering procedures. This paper presents a wavelet-based multiscale products thresholding and spatial based Bilateral Filter scheme for noise removal of magnetic resonance images. In the first stage, the noisy image is passed through Bilateral Filter (BF) and some amount of noise is reduced but the image becomes blurred, hence adaptive wavelet thresholding is applied with multiscale product rule in the second stage. Rather than using Wavelet Transform, Dyadic Wavelet Transform is employed. After apply the Dyadic Wavelet Transform, wavelet interscale dependencies is exploited by multiply the adjacent wavelet subbands to preserve edge structures while suppressing noise. An adaptive threshold is calculated and implemented on the multiscale products rather than on the wavelet coefficients. Hence to identify important features like edges, curves and textures the thresholding is applied to multiscale products. The proposed method for medical image denoising is performed through a recursive application of simple steps and demonstrated on a number of standard medical images. The results obtained with the proposed method outperform other methods both visually and in terms of objective quality peak-signal-to-noise ratio (PSNR).

To the Editor: We read with interest the recent article by Elvan et al in Circulation: Arrhythmia and Electrophysiology .1 They compared the reliability of time-domain versus frequency-domain methods in determining the local activation rate in patients during atrial fibrillation (AF).2 The study provided potentially important information regarding the analysis of unipolar versus bipolar recordings. Unfortunately, essential details regarding the methods used for both the time-domain and the frequency-domain analyses were not presented. For example, information on the noise and frequency rejection criteria used for determination of AF cycle length in the time domain was nowhere to be found. Also crucial information on the …

In order to test the effect of noise on the various parameters of the SAECG, 83 patients underwent three consecutive recordings at different noise levels. The high noise (HN) recordings had a noise level of 0.60-0.74 microV, the intermediate noise (IN) had 0.31-0.59 microV, and the low noise (LN) had < or = 0.30 microV. For the calculation of noise we used the standard deviation of the mean noise of the composite lead high pass filtered at 40 Hz. The recordings were compared using time-domain, frequency-domain, and spectrotemporal analysis. The time-domain parameters of the LN recordings, using 25-Hz, 40-Hz, and 60-Hz high pass cutoffs, were significantly different from those of the HN or IN recordings (P < 0.05). In the frequency-domain analysis, significant differences were found in some of the parameters of the LN compared to the HN. The spectrotemporal analysis of the X and Z leads also showed significant differences among the LN and the other recordings. In the time-domain analysis, both at 40 Hz and 25 Hz, there were more abnormal LN compared to the HN recordings (P < 0.05). In the spectrotemporal analysis, there were significantly more abnormal HN and IN recordings compared to the LN (P < 0.001 and P < 0.01, respectively). Therefore, the level of noise, even within the acceptable range, can significantly affect the SAECG. In the time domain at the lower noise levels the parameters become more abnormal, while the opposite seems to occur in the spectral and the spectrotemporal analysis.

Heart rate variability (HRV) is a physiological phenomenon of the variation of a cardiac interval (interbeat) over time that reflects the activity of the autonomic nervous system. HRV analysis is usually based on electrocardiograms (ECG signals) and has found many applications in the diagnosis of cardiac diseases, including valvular diseases. This analysis could also be performed on seismocardiograms (SCG signals) and gyrocardiograms (GCG signals) that provide information on cardiac cycles and the state of heart valves. In our study, we sought to evaluate the influence of valvular heart disease on the correlations between HRV indices obtained from electrocardiograms, seismocardiograms, and gyrocardiograms and to compare the HRV indices obtained from the three aforementioned cardiac signals. The results of HRV analysis in the time domain and frequency domain of the ECG, SCG, and GCG signals are within the standard deviation and have a strong linear correlation. This means that despite the influence of VHDs on the SCG and GCG waveforms, the HRV indices are valid. Clinical Relevance-Cardiac mechanical signals (seismocar-diograms and gyrocardiograms) can be applied to evaluate heart rate variability despite the influence of valvular diseases on the morphology of cardiac mechanical signals.

Appropriate QRS detection of the ECG signal is extremely important for determining the variations in the heart rate. The two most famous algorithms for the QRS detection are Murthy and Rangaraj and Pan and Tompkin. Using these two algorithms, comparison of results for five different datasets i.e., MIT-BIH arrhythmia database, European ST-T database, the MIT-BIH Noise stress test database, the BIDMC congestive heart failure database and intracardiac atrial fibrillation database was done in MATLAB and comparative analysis was obtained on the basis of age, sex and heart rate (above 95 and below 60). Murthy and Rangaraj algorithm is the methodology which is derivative-based and provides results without any delay and gives reduced noise and sharp peak output. On the other hand, Pan and Tompkin algorithm performs the analysis of slope, width and amplitude of QRS complexes and is found to have higher accuracy than other real-time methods.

Respiration Rate (RR) is an important physiological indicator and plays a major role in health deterioration monitoring. Despite that, it has been neglected in hospital wards due to inadequate nursing skills and insufficient equipment. ECG signal, which is always monitored in a clinical setting, is modulated by respiration which renders it a highly enticing mean for the automatic RR estimation. In addition, accurate QRS detection is pivotal to RR estimation from the ECG signal. The investigation of QRS complexes is a continuing concern in ECG analysis because current methods are still inaccurate and miss heart beats. This paper presents a frequency domain RR estimation method which uses a novel real-time QRS detector based on Empirical Mode Decomposition (EMD). Another novelty of the proposed work stems from the RR estimation in the frequency domain as opposed to some of the current methods which rely on a time domain analysis. As will be shown later, the RR extraction in the frequency domain provides more accurate results compared to the time domain methods. Moreover, our novel QRS detector uses an adaptive threshold over a sliding window and differentiates large Q- from R-peaks, facilitating a more accurate RR estimation. The performance of our methods was tested on real data from Capnobase dataset. An average mean absolute error of less than 0.5 breath per minute was achieved using our frequency domain method, compared to 6 breaths per minute of the time domain analysis. Moreover, our modified QRS detector shows comparable results to other published methods. achieving a detection rate over 99.80%.

Adaptive filter is a primary method to filter ECG signal, because it does not need the signal statistical characteristics. In this paper we present a novel adaptive filter for removing the artifacts from ECG signals based on Constrained Stability Least Mean Square (CSLMS) algorithm. This algorithm is derived based on the minimization of the squared Euclidean norm of the difference weight vector under a stability constraint defined over the posteriori estimation error. The adaptive filter essentially minimizes the mean-squared error between a primary input, which is the noisy ECG, and a reference input, which is either noise that is correlated in some way with the noise in the primary input or a signal that is correlated only with ECG in the primary input. Different filter structures are presented to eliminate the diverse forms of noise. Finally, we have applied this algorithm on ECG signals from the MIT-BIH data base and compared its performance with the conventional LMS algorithm. The results show that the performance of the CSLMS based algorithm is superior to that of the LMS based algorithm in noise reduction.

In this paper, a wearable belt-type ECG electrode worn around the chest by measuring the real-time ECG is produced in order to minimize the inconvenient in wearing. ECG signal is detected using a potential instrument system. The measured ECG signal is transmits via an ultra low power consumption wireless data communications unit to personal computer using Zigbee-compatible wireless sensor node. ECG signals carry a lot of clinical information for a cardiologist especially the R-peak detection in ECG. R-peak detection generally uses the threshold value which is fixed. There will be errors in peak detection when the baseline changes due to motion artifacts and signal size changes. Preprocessing process which includes differentiation process and Hilbert transform is used as signal preprocessing algorithm. Thereafter, variable threshold method is used to detect the R-peak which is more accurate and efficient than fixed threshold value method. R-peak detection using MIT-BIH databases and Long Term Real-Time ECG is performed in this research in order to evaluate the performance analysis.

The unsupervised pattern recognition for the ECG signal features detection