Detection of ECG characteristic points using wavelet transforms

The use of the Hilbert transform in ECG signal analysis

This paper presents a new robust algorithm for QRS detection using the first differential of the ECG signal and its Hilbert transformed data to locate the R wave peaks in the ECG waveform. Using this method, the differentiation of R waves from large, peaked T and P waves is achieved with a high degree of accuracy. In addition, problems with baseline drift, motion artifacts and muscular noise are minimised. The performance of the algorithm was tested using standard ECG waveform records from the MIT-BITH Arrhythmia database. An average detection rate of 99.87%, a sensitivity (Se) of 99.94% and a positive prediction (+P) of 99.93% have been achieved against study records from the MIT-BITH Arrhythmia database. A detection error rate of less than 0.8% was achieved in every study case. The reliability of the proposed detector compares very favorably with published results for other QRS detectors.

A hybrid on-line ECG segmenting system for long-term monitoring

This paper introduces a new hybrid ECG beat segmenting system, which can be applied in the processing unit of single-channel, long-term ECG monitors for the on-line segmentation of the ECG signal. Numerous ECG segmentation techniques are already existing and applied, however sufficiently robust and reliable methods currently require more than one ECG signal channel and quite complex computations, which are practically not feasible in stand-alone, low-cost monitors. Our new system approach presents a time domain segmentation technique based on a priori physiological and morphological information of the ECG beat. The segmentation is carried out after classifying the ECG beat, using the linear approximation of the filtered ECG signal and considering the pathophysiological properties as well. The proposed algorithms require moderate computational power, allowing the practical realization in battery powered stand-alone long-term cardiac monitors or small-sized cardiac defibrillators. The prototype version of the system was implemented in Matlab. The test and evaluation of the system was carried out with the help of reference signal databases.

Automatic detection of rapid eye movements by discrete wavelet transform

In order to detect rapid eye movements (REM) automatically, the Discrete Wavelet Transform was applied to each 8-s segment of electrooculogram (EOG) data for 30 min of 8 h of normal sleep. The Haar function was used as an analysing wavelet because this function is similar to the REM waveform. By shifting the phase of the analysing wavelet by pi/4 of the function, 96% of REM could be detected. The artifacts caused by body movements could be detected simultaneously by this method. Computing time required for the detection of REM was only 11 s for 30 min EOG data.

Analysis of synaptic quantal depolarizations in smooth muscle using the wavelet transform

The time-frequency characteristics of synaptic potentials contain valuable information about the process of neurotransmission between nerves and their target organs. For example, at the synapse between autonomic nerves and smooth muscle, two central issues of neurophysiology, i.e., 1) the probability of neurotransmitter release and 2) the quantal behavior of transmission can be deduced from analysis of the rising phases of evoked excitatory junction potentials (eEJP's) recorded from smooth muscle. eEJP rising phases are marked by prominent inflexions, which reflect these features of neuronal activity. Since these inflexions contain time-varying frequency information, we have applied recent techniques of time-frequency analysis based upon wavelet transforms to eEJP's recorded from the guinea-pig vas deferens in vitro. We find that these techniques allow accurate and convenient characterization of neuronal release sites, and that their probability of release falls between 0.001-0.004. We have also analyzed eEJP's recorded in the presence of the chemical 1-heptanol, which reveals quantal depolarizations. These results have helped clarify the nature of the quantal depolarizations that underly eEJP's. The present method offers significant advantages over those previously employed for these tasks, and holds promise as a novel approach to the analysis of synaptic potentials.

Uterine EMG analysis: a dynamic approach for change detection and classification

Toward the goal of detecting preterm birth by characterizing events in the uterine electromyogram (EMG), we propose a method of detection and classification of events in this signal. Uterine EMG is considered as a nonstationary signal and our approach consists of assuming piecewise stationarity and using a dynamic change detector with no a priori knowledge of the parameters of the hypotheses on the process state to be detected. The detection approach is based on the dynamic cumulative sum (DCS) of the local generalized likelihood ratios associated with a multiscale decomposition using wavelet transform. This combination of DCS and multiscale decomposition was shown to be very efficient for detection of both frequency and energy changes. An unsupervised classification based on the comparison between variance-covariance matrices computed from selected scales of the decomposition was implemented after detection. Finally a class labeling based on neural networks was developed. This algorithm of detection-classification-labeling gives satisfactory results on uterine EMG: in most cases more than 80% of the events are correctly detected and classified whatever the term of gestation.

A new method for the extraction of fetal ECG from the composite abdominal signal

We developed a wavelet transform-based method to extract the fetal electrocardiogram (ECG) from the composite abdominal signal. This is based on the detection of the singularities obtained from the composite abdominal signal, using the modulus maxima in the wavelet domain. Modulus maxima locations of the abdominal signal are used to discriminate between maternal and fetal ECG signals. Two different approaches have been considered. In the first approach, at least one thoracic signal is used as the a prior to perform the classification whereas in the second approach no thoracic signal is needed. A reconstruction method is utilized to obtain the fetal ECG signal from the detected fetal modulus maxima. The proposed technique is different from the classical time-domain methods, in that we exploit the most distinct features of the signal, leading to more robustness with respect to signal perturbations. Results of experiments with both synthetic and real ECG data have been presented to demonstrate the efficacy of the proposed method.

Multiscale characterization of chronobiological signals based on the discrete wavelet transform

To compensate for the deficiency of conventional frequency-domain or time-domain analysis, this paper presents a multiscale approach to characterize the chronobiological time series (CTS) based on a discrete wavelet transform (DWT). We have shown that the local modulus maxima and zero-crossings of the wavelet coefficients at different scales give a complete characterization of rhythmic activities. We further constructed a tree scheme to represent those interacting activities across scales. Using the bandpass filter property of the DWT in the frequency domain, we also characterized the band-related activities by calculating energy in respective rhythmic bands. Moreover, since there is a fast and easily implemented algorithm for the DWT, this new approach may simplify the signal processing and provide a more efficient and complete study of the temporal-frequency dynamics of the CTS. Preliminary results are presented using the proposed method on the locomotion of mice under altered lighting conditions, verifying its competency for CTS analysis.

A novel wavelet transform based analysis reveals hidden structure in ventricular fibrillation

We report a new method of interrogating the surface ECG signal using techniques developed in the field of wavelet transform analysis. Previously unreported structure within the ECG during ventricular fibrillation (VF) is found using a high-resolution decomposition of the signal employing the continuous wavelet transform. We believe that wavelet transform methods could lead to the development of powerful tools for use in the resuscitation of patients with cardiac arrest.

Feature extraction and quantification of the variability of dynamic performance profiles due to the different sagittal lift characteristics

Investigation of manual material handling (MMH) tasks, such as lifting, requires the quantification of the various kinematic and kinetic parameters of performance for assessment of the functional capacity and/or task demand profiles. Traditional statistical descriptive analyses usually involve computing the summary statistics (maximum, minimum, mean, and/or range) of the resulting performance parameters over the cycle duration (i.e., lifting/lowering cycle). Consequently, the significant information content of the time-varying signals is diminished, limiting the sensitivity of subsequent hypothesis testing procedures. The present study developed a methodology for representing and quantifying performance data variability of the kinematic and kinetic motion profiles due to the different lift characteristics (load, mode, and speed) during MMH tasks while capturing the temporal characteristics. Using a database of motion profiles from a manual lifting experiment, the Karhunen-Loeve Expansion (KLE) feature extraction technique was shown to be quite effective for representing the various motion profiles. The number of basis vectors (eigenvectors) and corresponding coefficients needed for accurate representation were substantially smaller than the original data set, resulting in data compression. Moreover, the effects of lift characteristics were investigated using analysis of variance techniques that recognize the vectorial constitution of the waveforms. The application of these techniques will enable the quantification of highly phasic profiles and enhance the ability to document the effect of intervening measures such as educational or physical training/exercise on the kinematic and kinetic patterns of performance. Additionally, the differential influence of lift characteristics on the variability of performance during different phases of lifting and lowering provides added resolution in the analysis of MMH tasks.

Novel real-time R-wave detection algorithm based on the vectorcardiogram for accurate gated magnetic resonance acquisitions

Electrocardiograph (ECG) triggered or gated magnetic resonance methods are used in many imaging applications. Therefore, a reliable trigger signal derived from to the R-wave of the ECG is essential, especially in cardiac imaging. However, currently available methods often fail mainly due to the artifacts in the ECG generated by the MR scanner itself, such as the magnetohydrodynamic effect and gradient switching noise. The purpose this study was to characterize the accuracy of selected R-wave detection algorithms in an MR environment, and to develop novel approaches to eliminate imprecise triggering. Vectorcardiograms (VCG) in 12 healthy volunteers exposed to 1.5 T magnetic field were digitized and used as a reference data set including manually corrected onsets of R-waves. To define the magnetohydrodynamic effect, the VCGs were characterized in time, frequency, and spatial domains. The selected real-time R-wave detection algorithms, and a new "target-distance" VCG-based algorithm were applied either to standard surface leads calculated from the recorded VCG or to the VCG directly. The flow related artifact was higher in amplitude than the R-wave in 28% of the investigated VCGs which yielded up to 9-16%false positive detected QRS complexes for traditional algorithms. The "target-distance" R-wave detection algorithm yielded a score of 100% for detection with 0.2% false positives and was superior to all the other selected methods. Thus, the VCG of subjects exposed to a strong magnetic field can be use to separate the magnetohydrodynamic artifact and the actual R-wave, and markedly improves the trigger accuracy in gated magnetic resonance scans. Magn Reson Med 42:361-370, 1999.

Wavelet transform-based QRS complex detector

In this paper, we describe a QRS complex detector based on the dyadic wavelet transform (Dy WT) which is robust to time-varying QRS complex morphology and to noise. We design a spline wavelet that is suitable for QRS detection. The scales of this wavelet are chosen based on the spectral characteristics of the electrocardiogram (ECG) signal. We illustrate the performance of the Dy WT-based QRS detector by considering problematic ECG signals from the American Heart Association (AHA) data base. Seventy hours of data was considered. We also compare the performance of Dy WT-based QRS detector with detectors based on Okada, Hamilton-Tompkins, and multiplication of the backward difference algorithms. From the comparison, results we observed that although no one algorithm exhibited superior performance in all situations, the Dy WT-based detector compared well with the standard techniques. For multiform premature ventricular contractions, bigeminy, and couplets tapes, the Dy WT-based detector exhibited excellent performance.

Biorthogonal wavelet transforms for ECG parameters estimation

The parameters of various morphologies of ECG waveform are basic in characterizing them as normal or otherwise. The use of multiscale analysis, through biorthogonal wavelets presented in this paper, appears very promising for such a characterization. This is on account of the fact that various morphologies are excited better at different scales. From these different scales, amplitudes, durations and various segments, widths can be determined more accurately. Simulation studies, with real ECG data, have shown that even when the signal-to-noise ratios are poor, the proposed technique can be used to accurately estimate the said parameters.

SVD-based on-line exercise ECG signal orthogonalization

An orthogonalization method to eliminate unwanted signal components in standard 12-lead exercise electrocardiograms (ECG's) is presented in this work. A singular-value-decomposition-based algorithm is proposed to decompose the signal into two time-orthogonal subspaces; one containing the ECG and the other containing artifacts like baseline wander and electromyogram. The method makes use of redundancy in 12-lead ECG. The same method is also tested for reconstruction of a completely lost channel. The online implementation of the method is given. It is observed that the first two decomposed channels with highest energy are sufficient to reconstruct the ST-segment and J-point. The dimension of the signal space, on the other hand, does not exceed three. Data from 23 patients, with duration ranging from 9 to 21 min, are used.

ECG beat detection using filter banks

We have designed a multirate digital signal processing algorithm to detect heart beats in the electrocardiogram (ECG). The algorithm incorporates a filter bank (FB) which decomposes the ECG into subbands with uniform frequency bandwidths. The FB-based algorithm enables independent time and frequency analysis to be performed on a signal. Features computed from a set of the subbands and a heuristic detection strategy are used to fuse decisions from multiple one-channel beat detection algorithms. The overall beat detection algorithm has a sensitivity of 99.59% and a positive predictivity of 99.56% against the MIT/BIH database. Furthermore this is a real-time algorithm since its beat detection latency is minimal. The FB-based beat detection algorithm also inherently lends itself to a computationally efficient structure since the detection logic operates at the subband rate. The FB-based structure is potentially useful for performing multiple ECG processing tasks using one set of preprocessing filters.

Adaptive denoising and multiscale detection of the V wave in brainstem auditory evoked potentials

This paper describes a wavelet-transform-based system for the V wave identification in brainstem auditory evoked potentials (BAEP). The system combines signal denoising and rule-based localization modules. The signal denoising module has the potential of effective noise reduction after signal averaging. It analyses adaptively the evolution of the wavelet transform maxima across scales. The singularities of the signal create wavelet maxima with different properties from those of the induced noise. A non-linear filtering process implemented with a neural network extracts out the noise-induced maxima. The filtered wavelet details are subsequently analysed by the rule-based localization module for the automatic identification of the V wave. In the first phase, it implements a set of statistical observations as well as heuristic criteria used by human experts in order to classify the IV-V complex. At the second phase, using a multiscale focusing algorithm, the IV and V waves are positioned on the BAEP signal. Our experiments revealed that the system provides accurate results even for signals exhibiting unclear IV-V complexes.

ECG pattern recognition and classification using non-linear transformations and neural networks: a review

The most widely used signal in clinical practice is the ECG. ECG conveys information regarding the electrical function of the heart, by altering the shape of its constituent waves, namely the P, QRS, and T waves. Thus, the required tasks of ECG processing are the reliable recognition of these waves, and the accurate measurement of clinically important parameters measured from the temporal distribution of the ECG constituent waves. In this paper, we shall review some current trends on ECG pattern recognition. In particular, we shall review non-linear transformations of the ECG, the use of principal component analysis (linear and non-linear), ways to map the transformed data into n-dimensional spaces, and the use of neural networks (NN) based techniques for ECG pattern recognition and classification. The problems we shall deal with are the QRS/PVC recognition and classification, the recognition of ischemic beats and episodes, and the detection of atrial fibrillation. Finally, a generalised approach to the classification problems in n-dimensional spaces will be presented using among others NN, radial basis function networks (RBFN) and non-linear principal component analysis (NLPCA) techniques. The performance measures of the sensitivity and specificity of these algorithms will also be presented using as training and testing data sets from the MIT-BIH and the European ST-T databases.

Application of adaptive signal processing for determining the limits of P and T waves in an ECG

A new algorithm for the determination of the limits of P and T waves is proposed, and its foundations are mathematically analyzed. The algorithm performs an adaptive filtering so that the searched point corresponds to a minimum. Crucial properties of its performance are discussed, i.e., immunity to base line drifts and full adaptation to any cardiological criteria. A series of tests are made involving real registers with different morphologies for P and T-waves.

Application of the matching pursuit method for structural decomposition and averaging of phonocardiographic signals

The paper evaluates the performance of an automatic adaptive time-frequency method to detect each cardiac cycle of a phonocardiogram (PCG) and extract average heart sounds and PCG cycles. The proposed method combines a global search of the PCG, in terms of the energy distribution of the most important components, with a local search relating to the specific events found within a cardiac cycle. The method is applied to 100 PCG recordings from 50 patients with an aortic bioprosthetic valve. The performance of the proposed method is compared with a commonly used semi-automatic method that is based on the combined analysis of an electrocardiogram (ECG) and the PCG signal. Results show that the proposed method clearly outperforms the semi-automatic method, especially in the case of patients with malfunctioning bioprostheses. By eliminating the need to record an ECG as the time-reference signal, this method reduces hardware overheads when analysis of PCG signals is the primary aim. It is also independent of subjective human judgment for selection of reference templates and threshold levels. Furthermore, the method is robust to artefacts, background noise and other kinds of signal interferences. With minor modifications, the procedure described could be applied to other types of biomedical signal in order to extract coherent transient components and identify specific events.

Heart response to horizontal impulse

Lumped body parameters linear and nonlinear models have been developed and used for the analysis of the response of the heart in a seated human body due to impulsive horizontal inputs at various body segments. The acceleration transfer magnitude and phase due to impulsive inputs at various body segments are reported. Time histories of the heart acceleration transfer were obtained for both linear and nonlinear models. The results indicate that the largest acceleration transfer occurs at 2-3 Hz frequency. Inputs at the upper body segments excite a second peak and in the acceleration transfer at 10-12 Hz. The nonlinear model shows large attenuation at the high frequency range (larger than 10 Hz) and less attenuation at the 1-5 Hz frequency range.

Wavelet analysis of embolic heart sound detected by precordial Doppler ultrasound during continuous venous air embolism in dogs

The spectrum of the embolic heart sounds (EHS) detected by precordial Doppler ultrasound has been previously characterized, but only on small volumes of venous air embolism (VAE). We sought to determine whether real-time wavelet analysis is useful in analyzing the signals of EHS and whether the embolic power of the EHS for larger volumes of air is proportionate to the volume of VAE that has been reported for small volumes of VAE. A series of small air boli (0.01, 0.02, 0.05, 0.07, 0.1, 0.15, 0.2, 0.3, 0.4, and 0.8 mL), followed by continuous infusion of larger volumes of air (0.8, 1.6, 2.4, 4.8, and 9.6 mL), was injected into the external jugular vein through a central catheter in seven pentobarbital-anesthetized dogs. We measured the spectrum of the Doppler heart sound (DHS) in a real-time manner by using wavelet analysis at different scales. Wavelet analysis at scale = 1 yielded satisfactory results in distinguishing abnormal EHS from normal DHS with high sensitivity (100%) and good positive predictive value (100%) compared with the conventional method, which requires an anesthesiologist to listen to the audio DHS signals in a real-time manner. There was a linear relationship (y = 1.08x + 7.89, r = 0.75, P < 0.001) between the cumulative embolic power of the EHS and the air volume introduced in the form of either bolus or continuous infusion. The 95% confidence intervals for slope and intercept were 0.89-1.27 and 7.65-8.13, respectively. Our results suggest that wavelet analysis is effective as a real-time monitor and that it is possible to distinguish larger volumes of air emboli based on previous injections of small volumes of air.

IMPLICATIONS

The real-time wavelet analysis of the heart sound detected by precordial Doppler ultrasound may be useful in estimating larger volumes of air emboli based on previous injections of small volumes of air in anesthetized dogs.

Wavelet transform in electrocardiography--data compression

An application of the wavelet transform to electrocardiography is described in the paper. The transform is used as a first stage of a lossy compression algorithm for efficient coding of rest ECG signals. The proposed technique is based on the decomposition of the ECG signal into a set of basic functions covering the time-frequency domain. Thus, non-stationary character of ECG data is considered. Some of the time-frequency signal components are removed because of their low influence to signal characteristics. Resulting components are efficiently coded by quantization, composition into a sequence of coefficients and compression by a run-length coder and a entropic Huffman coder. The proposed wavelet-based compression algorithm can compress data to average code length about 1 bit/sample. The algorithm can be also implemented to a real-time processing system when wavelet transform is computed by fast linear filters described in the paper.

Denoising of the fetal heart rate signal with non-linear filtering of the wavelet transform maxima

The fetal heart rate (FHR) signal provides valuable information for fetal development and well-being. However, the FHR traces derived from present-day ultrasound cardiotocographs are not of the desired quality. The paper applies the wavelet transform (WT) in order to denoise effectively the FHR signal. The denoising procedure analyses the evolution of the WT maxima across scales. The singularities of the signal create wavelet maxima with different properties from those of the induced noise. Since it is difficult to formulate precise rules that distinguish between the wavelet maxima of the FHR signal from those of the noise we have trained a neural network for this classification task. The neural network draws out successfully the noise induced wavelet maxima. An improved FHR signal can be obtained from the coarser wavelet approximation signal component and the filtered wavelet maxima by means of the inverse dyadic wavelet transform. Also, feature extraction and processing algorithms can be defined on the denoised wavelet coefficients (instead of on the original signal).

Wavelet and wavelet packet compression of electrocardiograms

Wavelets and wavelet packets have recently emerged as powerful tools for signal compression. Wavelet and wavelet packet-based compression algorithms based on embedded zerotree wavelet (EZW) coding are developed for electrocardiogram (ECG) signals, and eight different wavelets are evaluated for their ability to compress Holter ECG data. Pilot data from a blind evaluation of compressed ECG's by cardiologists suggest that the clinically useful information present in original ECG signals is preserved by 8:1 compression, and in most cases 16:1 compressed ECG's are clinically useful.

DSP implementation of wavelet transform for real time ECG wave forms detection and heart rate analysis

An algorithm based on wavelet transform (WTs) suitable for real time implementation has been developed in order to detect ECG characteristics. In particular, QRS complexes, P and T waves may be distinguished from noise, baseline drift or artefacts. This algorithm is implemented in a DSP (SPROC-1400) with a 50 MHz frequency clock. The performance of this algorithm is discussed, its accuracy is evaluated and a comparison is made with a similar algorithm implemented in C language. For the standard MIT/BIH arrhythmia database, this algorithm correctly detects 99.7% of the QRS complexes.

Temporal and spatial analysis of potential maps via multiresolution decompositions

Cardiac potentials recorded on the epicardium or the body surface by an array of electrodes are usually analyzed either as spatial distributions or temporal waveforms. Thus, the analysis often involves temporal descriptors (eg. max dV/dt) or spatial descriptors (eg. location of local extrema) only. The best known transform technique that has been applied to these data that combines both spatial and temporal characteristics is the Karhunen-Loeve transform, a global transform applied to temporal and/or spatial bases obtained by statistical analysis of a database. As an alternative, multiresolution decompositions and related wavelet-type transforms have recently seen great development in signal processing and related fields. They offer flexibility, employing transformations onto local (rather than global) and fixed (rather than data-dependent) databases, and allow transformation of distributions, waveforms, or both, as desired. The utility of this method as applied to temporal and spatial segmentation and analysis of map data from both epicardial plaques and body surface potentials recorded during percutaneous transluminal coronary angioplasty is illustrated.