Wavelet based ST-segment analysis

A novel algorithm for ST-segment analysis is developed using the multi-resolution wavelet approach. The system detects the QRS complexes and analyses each beat using the wavelet transform to identify the characteristic points (fiducial points). These fiducial points are, iso-electric level, the J point, and onsets and offsets of the QRS complex and T wave. The algorithm determines the T onset by looking for a point of inflection between the J point and the T peak. Furthermore, detection of characteristic points by the wavelet technique reduces the effect of noise. The results show that the proposed approach gives very accurate ST levels, as compared to the conventional (empirical) technique, at higher heart rates and with different morphologies. The algorithm detects the ST-segment length in 92.3% beats with an error of 4 ms, and in 97.3% beats the error is within 8 ms. The algorithm has been implemented on a TMS320C25 based add-on DSP card connected to a PC to provide the on-line analysis and display of ST-segment data.     

K-nearest neighbour-based algorithm for P- and T-waves detection and delineation

The aim of an automated Electrocardiogram (ECG) delineation system is the reliable detection of the characteristic waveforms and determination of peaks and limits of individual QRS-complex, P- and T-waves. In this paper, a classical statistical pattern recognition algorithm characterized with high accuracy and stability, i.e., K-Nearest Neighbour (KNN) has been proposed for locating the fiducial points along with their waveform boundaries in ECG signals. First, the QRS-complex along with its onset and offset points of each beat is detected from the ECG signal. After that P- and T-wave, relative to each QRS-complex along with their onset and offset points, are then identified using this algorithm. The feature extraction is done using the gradient of the ECG signals. The performance of the proposed algorithm has been evaluated on two standard manually annotated databases, (i) CSE and (ii) QT, and also on ECG data acquired using BIOPAC®MP100 system in laboratory settings. The results in terms of accuracy, i.e., 92.8% for CSE database obtained, clearly indicate a high degree of agreement with the manual annotations made by the referees of CSE dataset-3. Further, the delineation results of the CSE and QT database are compared with the accepted tolerances as recommended by the CSE working party. The results for ECG records acquired using the BIOPAC®MP100 system, in terms of QRS duration, heart rate, QT-interval, P-wave duration and PR-interval using KNN algorithm have also been computed.     

Segmentation of Holter ECG Waves Via Analysis of a Discrete Wavelet-Derived Multiple Skewness–Kurtosis Based Metric

In this study, a simple mathematical-statistical based metric called Multiple Higher Order Moments (MHOM) is introduced enabling the electrocardiogram (ECG) detection–delineation algorithm to yield acceptable results in the cases of ambulatory holter ECG including strong noise, motion artifacts, and severe arrhythmia(s). In the MHOM measure, important geometric characteristics such as maximum value to minimum value ratio, area, extent of smoothness or being impulsive and distribution skewness degree (asymmetry), occult. In the proposed method, first three leads of high resolution 24-h holter data are extracted and preprocessed using Discrete Wavelet Transform (DWT). Next, a sample to sample sliding window is applied to preprocessed sequence and in each slid, mean value, variance, skewness, and kurtosis of the excerpted segment are superimposed called MHOM. The MHOM metric is then used as decision statistic to detect and delineate ECG events. To show advantages of the presented method, it is applied to MIT-BIH Arrhythmia Database, QT Database, and T-Wave Alternans Database and as a result, the average values of sensitivity and positive predictivity Se = 99.95% and P+ = 99.94% are obtained for the detection of QRS complexes, with the average maximum delineation error of 6.1, 4.1, and 6.5 ms for P-wave, QRS complex, and T-wave, respectively showing marginal improvement of detection–delineation performance. In the next step, the proposed method is applied to DAY hospital high resolution holter data (more than 1,500,000 beats including Bundle Branch Blocks—BBB, Premature Ventricular Complex—PVC, and Premature Atrial Complex—PAC) and average values of Se = 99.97% and P+ = 99.95% are obtained for QRS detection. In summary, marginal performance improvement of ECG events detection–delineation process, reliable robustness against strong noise, artifacts, and probable severe arrhythmia(s) of high resolution holter data can be mentioned as important merits and capabilities of the proposed algorithm.     

A new approach to QRS segmentation based on wavelet bases and adaptive threshold technique

In this paper, we develop and evaluate a new approach to QRS segmentation based on the combination of two techniques: wavelet bases and adaptive threshold. Firstly, QRS complexes are identified without a preprocessing stage. Then, each QRS is segmented by identifying the complex onset and offset. We evaluated the algorithm on two manually annotated databases, the QT-database and the MIT-BIH Arrhythmia database. The QRS detector obtained a sensitivity of 99.02% and a positive predictivity of 99.35% over the first lead of the validation databases (more than 192,000 beats), while for the QT-database, values larger than 99.6% were attained. As for the delineation of the QRS complex, the mean and the standard deviation of the differences between the automatic and the manual annotations were computed. Using QT-database which contains recordings of annotated ECG with a sampling rate of 250 Hz, we obtain the average of the differences not exceeding two sampling intervals, while the standard deviations were within acceptable range of values.     

Combining neural networks and ANFIS classifiers for supervised examining of electrocardiogram beats

Abstract

In this paper, the supervised classification of the electrocardiogram (ECG) beats based on the fusion of several intelligent learning machines is described. For classification of ECG heartbeats, first, the QRS complexes are delineated by an efficient algorithm so as to identify the fiducial and J-locations of each complex. For each delineated QRS complex, a feature vector is established based on the geometrical properties of the complex waveform and its associated discrete-wavelet transform. Next, three different multi-layer perceptron back-propagation (MLP-BP) networks are trained with different topologies and intrinsic parameters. Afterwards, the outputs of MLP-BPs are used as the new feature space elements for training three adaptive fuzzy network inference systems (ANFIS) in order to increase the final accuracy. At the end, the outputs of ANFIS classifiers are voted based on majority for each input sample. The method was applied to seven arrhythmias (Normal, LBBB, RBBB, PVC, APB, VE, VF) which belong to the MIT-BIH Arrhythmia Database and the average accuracy value Acc?=?98.28% was achieved for the beat-level. Also, the proposed method was assessed to five arrhythmias (Normal, LBBB, RBBB, PVC, APB) according to validation standards of the American Heart Association (AHA) at record (subject) level and the average accuracy value Acc?=?73.39% was achieved. To evaluate performance quality of the new proposed hybrid learning machine, the obtained results were compared with similar peer-reviewed studies in this area.     

New algorithm for the detection of the ECG fiducial point in the averaging technique

The use of the coherent averaging technique applied to the electrocardiographic signal implies the location of a fiducial point as a synchronisation reference. An algorithm easily adaptable to a personal computer, operable in real time, insensitive to mains and to ECG-baseline fluctuations, with a low jitter value and the capacity to trigger any ECG signal wave or complex, has been developed. The algorithm detects those waveforms which, within certain confidence intervals, are morphologically equal to a reference wave. This wave is chosen by the user as the repetitive waveform within which the fiducial point is to be located. A two-window template and differential parameters are used. The possibility of building the template permits the user to adapt the algorithm to each patient's ECG. To evaluate its accuracy objectively, a software simulation was built of a generator capable of producing test signals as the sum of the 'useful' signal plus 'noise'. A jitter standard deviation of 1.65 ms was obtained in the worst test (SNR = 10 dB; noiseband = 0-50 Hz), which shows the excellent recognition accuracy of the algorithm.     

An innovative approach of QRS segmentation based on first-derivative, Hilbert and Wavelet Transforms

The QRS detection and segmentation processes constitute the first stages of a greater process, e.g., electrocardiogram (ECG) feature extraction. Their accuracy is a prerequisite to a satisfactory performance of the P and T wave segmentation, and also to the reliability of the heart rate variability analysis. This work presents an innovative approach of QRS detection and segmentation and the detailed results of the proposed algorithm based on First-Derivative, Hilbert and Wavelet Transforms, adaptive threshold and an approach of surface indicator. The method combines the adaptive threshold, Hilbert and Wavelet Transforms techniques, avoiding the whole ECG signal preprocessing. After each QRS detection, the computation of an indicator related to the area covered by the QRS complex envelope provides the detection of the QRS onset and offset. The QRS detection proposed technique is evaluated based on the well-known MIT-BIH Arrhythmia and QT databases, obtaining the average sensitivity of 99.15% and the positive predictability of 99.18% for the first database, and 99.75% and 99.65%, respectively, for the second one. The QRS segmentation approach is evaluated on the annotated QT database with the average segmentation errors of 2.85±9.90ms and 2.83±12.26ms for QRS onset and offset, respectively. Those results demonstrate the accuracy of the developed algorithm for a wide variety of QRS morphology and the adaptation of the algorithm parameters to the existing QRS morphological variations within a single record.     

Sources of common computer-cardiologist discord in e.c.g. interpretations in ambulatory population

Based on clinical experience in performing mass computerised interpretations of e.c.g.s, some common sources of computer-cardiologist disagreements are presented. Examples are given along with a discussion of the technological or logical bases for discord. It is concluded that computer interpretations are generally reliable and accurate; however, disagreements sometimes occur in borderline cases, in tracings with much electrical ‘noise’ or in those where the transitions between different segments of the QRS complex are not well demarcated. A computerised e.c.g. interpretation system can help the cardiologist by providing all required measurements and most of the diagnostic statements. Thus, only review and validation of the processed tracings are needed. By continuously refining the program criteria based on the accumulated experience of many cardiologists, the reliability and acceptability of computer e.c.g. interpretations could be enhanced and expanded to screen large populations.     

Evaluation of a new e.c.g. measurement program for the detection of rhythm disturbances

A new computer program has been developed to extract pertinent parameters from the adult e.c.g., for the purpose of automatic arrhythmia interpretation in the context of routine analysis. No interpretation program would be acceptable by cardiologists without a rhythm interpretation coupled to the morphology interpretation even if in routine situations rhythm interpretation is not as stringent as in monitoring environments. The measurements program operates simultaneously on the X, Y and Z leads of the modified Frank system. The e.c.g. features considered are the QRS wavetrain, the QRS morphological characteristics and the P wavetrain. Although the program is still under evaluation, a number of important results have been obtained based on the analysis of a given population of e.c.g.s. The performance on QRS detection gives a total error of 0·79% when adding the number of omitted QRS complexes and false det ections, and the system is able to realise perfect detection on over 95% of the e.c.g.s analysed. The performance on QRS classification is 1·85% of misclassified waveforms, with a further 0·67% which remains undefined due to the limitation in measurement capability of the algorithm used; 89·7% of the tracings are completely error free. The identification of the P wavetrain is achieved with a good degree of success. On a beat-to-beat basis, the program finds 92·5% of the P waves present in the tracings while generating 2·2% extraneous events. However, only 53% of the tracings are error free.     

A robust wavelet-based multi-lead electrocardiogram delineation algorithm

A robust multi-lead ECG wave detection-delineation algorithm is developed in this study on the basis of discrete wavelet transform (DWT). By applying a new simple approach to a selected scale obtained from DWT, this method is capable of detecting QRS complex, P-wave and T-wave as well as determining parameters such as start time, end time, and wave sign (upward or downward). First, a window with a specific length is slid sample to sample on the selected scale and the curve length in each window is multiplied by the area under the absolute value of the curve. In the next step, a variable thresholding criterion is designed for the resulted signal. The presented algorithm is applied to various databases including MIT-BIH arrhythmia database, European ST-T Database, QT Database, CinC Challenge 2008 Database as well as high resolution Holter data of DAY Hospital. As a result, the average values of sensitivity and positive predictivity Se = 99.84% and P+ = 99.80% were obtained for the detection of QRS complexes, with the average maximum delineation error of 13.7 ms, 11.3 ms and 14.0 ms for P-wave, QRS complex and T-wave, respectively. The presented algorithm has considerable capability in cases of low signal-to-noise ratio, high baseline wander, and abnormal morphologies. Especially, the high capability of the algorithm in the detection of the critical points of the ECG signal, i.e. the beginning and end of T-wave and the end of the QRS complex was validated by cardiologists in DAY hospital and the maximum values of 16.4 ms and 15.9 ms were achieved as absolute offset error of localization, respectively.     

QRS wave detection

A QRS complex detector based on optimum predetection with a matched filter is described. In order to improve the accuracy of the QRS complex recognition under conditions of Gaussian noise and variable QRS amplitude, the first derivative of the e.c.g. was used with zero threshold detection. In addition, two nonlinear circuits cut off low amplitude noise and all spikes which appear for a fixed time after QRS detection. Calculation of errors shows that differentiation reduces Gaussian error by √6 and errors caused by variable QRS amplitudes are close to zero. This detector is especially useful with biotelemetry systems since it reduces many interferences due to patient movement and communication channel distortion.     

An accurate programmable ECG simulator.

This article reports the design and development of an ECG simulator intended for use in the testing, calibration and maintenance of electrocardiographic equipment. It generates a lead II signal having a profile that varies with heart rate in a manner which reflects the true in vivo variation. Facilities are provided for user adjustment of heart rate, signal amplitude, QRS complex up-slope, and the relative amplitudes of the P-wave and T-wave. The heart rate can be set within the range 30-200 beats min(-1) in steps of 1 beat min(-1). The amplitude of the QR5 complex can be adjusted from 0.1-20 mV in 0.1 mV steps, while its up-slope can be set between 10 and 50 ms with a 1 ms resolution. The amplitude of the P-wave can be varied from 5-40% and that of the T-wave from 10-80% of the amplitude of the QRS complex with a 1% resolution.     

Fast and reliable QRS alignment technique for high-frequency analysis of signal-averaged ECG

The process of QRS alignment as required in signal-averaged ECG can impose serious limitations on the spectral range of the signal output. This effect depends basically on the particular alignment technique being used and on the level and type of noise present in the recorded ECG. In clinical studies where a wide-band (1000 Hz) ECG averager is required, the conventional QRS alignment technique, based on maximum coherence matching (MCM) with a template beat, may not perform consistently well. An alternative QRS alignment technique based on the accurate detection of a single fiducial point (SFP) in the bandpass filtered (3–30 Hz) QRS complex was developed. Using computer simulation methods, a comparative assessment of the frequency bandwidths (3 dB points) offered by both MCM and SFP techniques as a function of noise level (15–100 μ RMS) and type (EMG and 50 Hz interference), was carried out. The results of the comparative assessment indicated a better performance by the SFP technique in all cases of noise. Hence, the SFP technique would perform more reliably for high-frequency analysis of a noisy ECG, especially when 50 Hz interference is high. Furthermore, SFP is considerably faster than MCM (about four times) when implemented digitially, and its analogue realisation is feasible. The SFP technique is suitable for late-potential analysis in the signal-averaged ECG.     

Modeling of the heart's ventricular conduction system using fractal geometry: Spectral analysis of the QRS complex

Many biological systems having one or more characteristics that remain constant over a wide range of scales may be considered self-similar or fractal. Geometrical and functional overview of the ventricular conduction system of the heart reveals that it shares structures common to a tree with repeatedly bifurcating “branches,” decreasing in length with each generation. This system may further simplify by assuming that the bifurcating and decreasing process is the same at any generation, that is, the shortening factor and the angle of bifurcation are the same for each generation. Under these assumptions, the conduction system can be described as a fractal tree. A model of the heart's ventricles which consists of muscle cells and a fractal conduction system is described. The model is activated and the dipole potential generated by adjacent activated and resting cells is calculated to obtain a QRS complex. Analysis of the frequency spectrum of the QRS complex reveals that the simulated waveforms show an enhancement in the high frequency components as generations are added to the conduction system. It was also found that the QRS complex shows a form of an inverse power law, which was predicted by the fractal depolarization hypothesis, with a highly correlated straight line for a log-power versus log frequency plot with a slope of approximately −4. Similar results were obtained using real QRS data from healthy subjects.     

An accurate programmable ECG simulator

This article reports the design and development of an ECG simulator intended for use in the testing, calibration and maintenance of electrocardiographic equipment. It generates a lead II signal having a profile that varies with heart rate in a manner which reflects the true in vivo variation. Facilities are provided for user adjustment of heart rate, signal amplitude, QRS complex up-slope, and the relative amplitudes of the P-wave and T-wave. The heart rate can be set within the range 30 - 200 beats min^-1 in steps of 1 beat min^-1. The amplitude of the QRS complex can be adjusted from 0.1 - 20 mV in 0.1 mV steps, while its up-slope can be set between 10 and 50 ms with a 1 ms resolution. The amplitude of the P-wave can be varied from 5 - 40% and that of the T-wave from 10 - 80% of the amplitude of the QRS complex with a 1% resolution.     

Forward and inverse high-frequency electrocardiography

The high-frequency electrocardiogram (e.c.g.) contains useful clinical information about the heart condition that is not recorded in the conventional e.c.g. With simulated notches introduced in the heat dipole components and in the surface e.c.g., the forward and inverse problems of electrocardiography to correlate the standard 12-lead e.c.g. with the heart dipole were solved. It was shown that considerable notching activity in the heart dipole may be masked in the QRS complex of the body surface e.c.g. by large slopes or by cancellation effects. Furthermore, notches and slurs in the surface e.c.g. are observable even if the source (heart dipole) is notch-free. From theoretical analysis, the conclusion was reached that analysis of amplitudes and locations of notches and slurs in the vectorcardiogram can potentially be used clinically to obtain more information about the heart condition than the conventional technique of counting notches and slurs in the surface e.c.g. leads. The causes of notches are discussed and the sensitivity of each lead to notches is evaluated. Both homogeneous and inhomogeneous cases are examined.     

Cardiac vulnerability assessment from electrical microvariability of high-resolution electrocardiogram

Patients susceptible to malignant arrhythmias often have an increased beat-to-beat variation of the T-wave of the electrocardiogram. Variability analysis of the T-wave is increasingly used for non-invasive risk assessment. The aim of this study is to evaluate intra-QRS beat-to-beat signal variation and to compare it to ST-T variation. The beat-to-beat, microvolt variation of the QRS and the ST-T segment from 44 patients with coronary heart disease at high risk of suffering from malignant arrhythmias and from 51 healthy volunteers are compared. Variation analysis is carried out on 250 consecutive sinus beats from high-resolution electrocardiograms. The individual beats are filtered using a waveform-independent, cubic spline-filter. A variability index of the QRS and ST-T segments is calculated as the integrated standard deviation of corresponding samples inside the area of interest. Patients at risk of suffering from malignant arrhythmias have a significantly higher variability index of both the QRS (median 44.5 ms against 34.7 ms, p<0.001) and the ST-T segment (median 20.5 ms against 9.8 ms, p<0.001) compared to the group of healthy subjects. The discriminative ability of the odds variability indices of the QRS and ST-T segments are not statistically different, the ratios being 7.8 (QRS) and 12.6 (ST-T). We conclude that patients at high risk of suffering from malignant arrhythmias are characterised by an increased beat-to-beat microvolt variation of both the QRS and the ST-T segment. Further studies are necessary to evaluate the prognostic potential of depolarisation variability.