Massachusetts Institute Of Technology-Beth Israel Hospital Research Articles

Diagnostic quality assured ECG signal compression with selection of appropriate mother wavelet for minimal distortion

This study presents a diagnostic quality assured electrocardiogram (ECG) signal compression algorithm which uses discrete wavelet transform with the selection of appropriate mother wavelet. Since distortion of reconstructed ECG signal depends on the type of mother wavelet used for decomposition. Therefore, appropriate mother wavelet is selected first which produces minimal distortion. Small valued wavelet transform coefficients are discarded using dead-zone quantisation. Further integer conversion of coefficients is performed to improve compression at the cost of very less error. The processed transform coefficients obtained at this stage contain approximate coefficients and detail coefficients, where approximate coefficients consist very less repetition of data instances while detail coefficients are much repetitive. The repetition of detail coefficients is exploited by run-length encoding which represents the data as run and length. Compression performance of the proposed algorithm is evaluated using single-channel ECG records taken from the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia database, MIT-BIH ECG compression test database and Physikalisch-Technische Bundesanstalt (PTB) diagnostic ECG database. Compression results exhibit better performance than other existing techniques. Subjective evaluation of reconstructed ECG signals is also performed which ensures effective working of the proposed algorithm on the different morphology of ECG signals.

NI LabVIEW Biyomedikal Yazılımı ile EKG İşaretlerinin Öznitelik Çıkarımı ve Yapay Sinir Ağları ile Sınıflandırılması

In this study, a data set containing normal and different heart beat types recorded by the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) was used for the detection of cardiac dysfunctions. In this data set, features were extracted using the LabVIEW Biomedical Workbench from the normal heartbeat and six different arrhythmia types. Obtained signals were evaluated by using Artificial Neural Network multiple classification method. Classification performances were compared before extracting the feature on the same data set. Classifier performances were evaluated by accuracy, sensitivity and selectivity performances criteria of classification. In the classifier performances, the "Normal" beat rate was found to be 99% accurate with the highest success compared to other arrhythmia types. As a result, both analysis methods are successful, but when the LabVIEW Biomedical Workbench is used, the classification results have achieved higher success.

Hybrid approach for ECG signal enhancement using dictionary learning‐based sparse representation

Electrocardiogram (ECG) signal quality enhancement is a crucial task in the computer-based automated processing system. In this study, a dictionary learning (DL)-based sparse representation framework is presented for ECG signal enhancement through denoising. Unlike, traditional filtering techniques, the proposed method removes both low- and high-frequency noises for complete enhancement of the signal quality. The frequency localised DL-based sparse representation approach is applied to remove both baseline wander and power-line interference. The time-localised and signal characteristics based DL-based sparse representation scheme is employed for extraction of clean ECG components from white Gaussian noise and muscle artefact added noisy ECG record. Both qualitative and quantitative performance analyses are carried out using the Massachusetts Institute of Technology - Beth Israel Hospital (MIT-BIH) database, QT database and synthetic ECG signals. The efficacy of the proposed method is compared with the existing ECG denoising approaches using standard performance metrics: signal-to-noise ratio, root-mean-square error and percentage RMS difference. It is observed through the detailed study and analysis that the proposed approach outperforms the existing ones and can be served to further enhance the overall quality of ECG in a computer-based automated medical system.

Efficient FPGA‐based VLSI architecture for detecting R‐peaks in electrocardiogram signal by combining Shannon energy with Hilbert transform

Electrocardiogram (ECG) is a critical application in light of R-peak detection. The R-peaks are impacted by some QRS complex and noises in the current method. Basic testability, snappier execution, and confirmation choices are accomplished by the utilization of the field programmable gate array (FPGA). But, FPGA execution gives less accuracy. To battle execution trouble, another R-peak detector is proposed and relying upon the propelled peak finding logic, which incorporates a Bandpass filter and first-order differentiation process, which are done in the primary phase of the strategy. The noises in the input ECG signal are decreased in the first stage. In the subsequent stage, the smooth Shannon energy envelope (SEE) is acquired by utilizing SEE extraction and the zero phases filtering process. The false R peak is stifled by proposing a Hilbert transform (HT) in the third stage. The HT requires more hardware space and high-power utilization in the current technique. Due to this reason, Real valued Fast Fourier Transform (RFFT) and Inverse RFFT (IRFFT) techniques are proposed in the HT. In the fourth stage, R-peak acknowledgment is evaluated using Massachusetts Institute of Technology, Beth Israel Hospital (MIT-BIH) arrhythmia database and produced the average accuracy of 99.86%, sensitivity of 99.95% and the positive predictivity of 99.90%.

QP-DTW: Upgrading Dynamic Time Warping to Handle Quasi Periodic Time Series Alignment

Dynamic time warping (DTW) is the main algorithms for time series alignment. However, it is unsuitable for quasi-periodic time series. In the current situation, except the recently published the shape exchange algorithm (SEA) method and its derivatives, no other technique is able to handle alignment of this type of very complex time series. In this work, we propose a novel algorithm that combines the advantages of the SEA and the DTW methods. Our main contribution consists in the elevation of the DTW power of alignment from the lowest level (Class A, non-periodic time series) to the highest level (Class C, multiple-periods time series containing different number of periods each), according to the recent classification of time series alignment methods proposed by Boucheham (Int J Mach Learn Cybern, vol. 4, no. 5, pp. 537-550, 2013). The new method (quasi-periodic dynamic time warping [QP-DTW]) was compared to both SEA and DTW methods on electrocardiogram (ECG) time series, selected from the Massachusetts Institute of Technology - Beth Israel Hospital (MIT-BIH) public database and from the PTB Diagnostic ECG Database. Results show that the proposed algorithm is more effective than DTW and SEA in terms of alignment accuracy on both qualitative and quantitative levels. Therefore, QP-DTW would potentially be more suitable for many applications related to time series (e.g., data mining, pattern recognition, search/retrieval, motif discovery, classification, etc.).

Heartbeat classification with low computational cost using Hjorth parameters

A method for electrocardiogram (ECG) feature extraction is presented for automatic classification of heartbeats, using values of RR intervals, amplitude and Hjorth parameters. Hjorth parameters have been used in a variety of research areas, but their application to ECG signal processing is still little explored. This paper also introduces a new approach to heartbeat segmentation, which avoids mixing information from adjacent beats and improves classification performance. The proposed model is validated in the Massachusetts Institute of Technology - Beth Israel Hospital (MIT-BIH) Arrhythmia database and presents an overall accuracy of 90.4%, better than other state-of-the-art methods. There is an improvement over other models in positive predictivity for class S (66.6%) of supraventricular ectopic beats, and sensitivity for class N (93.0%). Results obtained indicate that the techniques used in this study can be successfully applied to the problem of automatic heartbeat classification. In addition, this new approach has low computational cost, which allows its later implementation in hardware devices with limited resources.

Cardiac Arrhythmias Automated Screening Using Discrete Fractional-Order Integration Process and Meta Learning Based Intelligent Classifier

Cardiac arrhythmias reveal multiple morphologies in time-domain waveforms. Abnormal beats reflect the origin and the conduction path of the ectopic heart activation pulses, including supraventricular, junctional, and ventricular arrhythmias and conduction abnormalities. This paper proposes a method for automated screening of cardiac arrhythmias using the discrete fractional-order integration (DFOI) and the meta-learning-based intelligent classifier. The DFOI process with the specific fractional order is used to extract the QRS features with the finite computations. It can deal with the irregular time-varying signals. The meta-learning-based intelligent classifier is used to identify the abnormal classes, consisting of a primary multilayer network and several agent networks. Since abnormal QRS complexes are multi-waveforms such as ventricular premature contraction, the classifier needs to retrain with new training patterns for meeting new condition in clinical applications. Thus, this model possesses the learning-to-optimization capability for retraining generalized regression neural network using the particle swarm optimization algorithm. Using incremental new training patterns, each inducer is used to track the past learning experiences through several agent networks. This pattern scheme can gradually enhance the accuracy by refining the optimal parameters of each sub-estimator for pattern recognition. Using the arrhythmia database of the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) database, the proposed intelligent classifier demonstrated greater efficiency and promising results in average accuracy, positive predictivity, and true negative rate for recognizing electrocardiography signals.

Efficient compression technique based on temporal modelling of ECG signal using principle component analysis

This study presents an improved technique for compression of electrocardiogram (ECG) signals, based on beat correlation of signal and principle component (PC) analysis, for ECG signal. For this purpose, two-dimensional matrix of ECG signal based on temporal inter-and intra-beat correlation is constructed, and further compression is achieved using PC extraction. Beat correlation helps to generate very few PCs that increase the compression efficiency. A detailed analysis has been presented for ten signals having different rhythms, wave morphologies and abnormalities of Massachusetts Institute of Technology - Beth Israel Hospital (MIT-BIH) arrhythmia database. The effectiveness of the proposed method is examined with several attributes such as percentage root-mean-square difference, compression ratio, signal-to-noise ratio and correlation. Experimental results have shown that this method is very efficient for compression and suitable for different applications of telecardiology.

ECG beat classification via deterministic learning

This paper proposes a novel method for the electrocardiographic (ECG) beat classification via deterministic learning. The dynamics of ECG beats is used as a unique feature for ECG beat classification, which is fundamentally different from the time/frequency domain features used in literature. It is the essential feature of ECG beats, and contains complete information of ECG beats. Precisely, the deterministic learning allows us to model and represent the dynamics of a training beat set as constant radial basis function (RBF) networks. As the classification measure, a set of errors is further obtained through the comparison between the test beat and the estimators constructed by the RBF networks. ECG records taken from the MIT-BIH (Massachusetts Institute of Technology-Beth Israel Hospital) arrhythmia database are selected to test the proposed method. With 5% beats used as training beats, the overall accuracies are 97.78% and 97.21% for global and patient-adapting beat classification, respectively. These results indicate the proposed method is reliable and efficient for ECG beat classification.

Heart beat classification from single-lead ECG using the synchrosqueezing transform

The processing of ECG signal provides a wealth of information on cardiac function and overall cardiovascular health. While multi-lead ECG recordings are often necessary for a proper assessment of cardiac rhythms, they are not always available or practical, for example in fetal ECG applications. Moreover, a wide range of small non-obtrusive single-lead ECG ambulatory monitoring devices are now available, from which heart rate variability (HRV) and other health-related metrics are derived. Proper beat detection and classification of abnormal rhythms is important for reliable HRV assessment and can be challenging in single-lead ECG monitoring devices. In this manuscript, we modelled the heart rate signal as an adaptive non-harmonic model and used the newly developed synchrosqueezing transform (SST) to characterize ECG patterns. We show how the proposed model can be used to enhance heart beat detection and classification between normal and abnormal rhythms. In particular, using the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia database and the Association for the Advancement of Medical Instrumentation (AAMI) beat classes, we trained and validated a support vector machine (SVM) classifier on a portion of the annotated beat database using the SST-derived instantaneous phase, the R-peak amplitudes and R-peak to R-peak interval durations, based on a single ECG lead. We obtained sentivities and positive predictive values comparable to other published algorithms using multiple leads and many more features.

FPGA system for QRS complex detection based on difference operation method

The detection of QRS complex is the start point for almost all electrocardiogram (ECG) applications. However, due to the complexity and non-causality of their mathematical computation, lots of QRS detectors can only be operated by using software but in a non-real-time fashion. This paper implements a real-time QRS complex detection system based on the difference operation method (DOM) and its efficient hardware architecture design. After simulations verified by MATLAB programs, we used the Verilog HDL (hardware description language) to perform the task of hardware design using the FPGA (field programmable gate array) system. The hardware design was tested with ECG records in the MIT-BIH (Massachusetts Institute of Technology- Beth Israel Hospital) arrhythmia database. The achievable accuracy exceeds 99% for records including typical and worst cases in the database.

Hardware implementation of ECG denoising system using TMS320C6713 DSP processor

The electrocardiogram (ECG) signal is an extensively used biomedical signal for diagnosis of heart diseases. However, the quality of ECG signal is deteriorated by several noises during its acquisition. The two dominant and recurring noises are power line interference and baseline wander noise and they have to be removed for better clinical evaluation. This paper proposes a new ECG denoising system using a combination of Empirical Mode Decomposition (EMD) algorithm and FFT-based frequency analysis. The proposed ECG denoising system is first simulated and validated using MATLAB. Then, it is implemented in TMS320C6713 DSP processor using Code Composer Studio (CCS). The proposed system is tested with the standard ECG signals obtained from Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) Arrhythmia Database. The root mean square error (RMSE) and correlation coefficient are used as the evaluation measure to compare the performance of the proposed method with an existing denoising system. The experimental results with reduced RMSE value and the average correlation coefficient of 0.9889 between the original ECG and the denoised one indicate the successful removal of power line interference and baseline wander noise by the proposed ECG denoising system.

Evolutionary optimisation of atrial fibrillation diagnostic algorithms

This research proposes evolutionary optimisation for the improvement of atrial fibrillation (AF) detection. The basis of the AF detection algorithm is the irregularity of RR intervals (heartbeats) in the electrocardiogram (ECG) signal. We use three well-known statistical methods to detect RR interval irregularity: root mean squares of successive differences (RMSSD), turning points ratio (TPR), and Shannon entropy (SE). The Massachusetts Institute of Technology/Beth Israel Hospital (MIT-BIH) AF databases are used in conjunction with biogeography-based optimisation (BBO) to tune the parameters of the statistical methods. We trained each of the three statistical methods to diagnose AF on each MIT-BIH database. Then we tested each trained detection method on the rest of the MIT-BIH databases. The accuracy levels that were achieved for the detection of AF were as high as 99% in the trained databases, and up to 75% in the tested databases. Compared to previously published AF detection accuracies, the proposed t...

An algorithm for detection of ECG morphological points using its non-linear energy

Analysis of an Electrocardiogram (ECG) signal for designing intelligent Clinical Decision Support System (CDSS) requires error-free and fast detection of ECG features. An algorithm dealing with Polarised Teager Energy (PTE) function has been proposed in this paper. According to the morphology of ECG, QRS complex contains highest non-linear energy. This property has been utilised for detection of R peak. Polarisation of Teager Energy (TE) yields remarkable decrease in false positive of R peaks and hence that of other morphological points. Results showed that the proposed algorithm comprises less computational overheads and still has a very good performance in QRS detection with 0.27% detection error rate, 99.85% sensitivity and 99.87% positive prediction rate. Detection error rate observed for the other morphological points ranges from 0.37 to 0.40. Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) database has been used for testing and validation of the algorithm.

Empirical Wavelet Transform Based ECG Signal Compression

ABSTRACTTransmission of biomedical signals over telephone lines or other communication channels is currently an important issue for the telemedicine applications. An efficient compression algorithm is needed to achieve a reduced information rate, for the storage and transmission purposes. In this paper, empirical wavelet transform (EWT) along with discrete wavelet transform (DWT) has been used for compression and reconstruction of the ECG signals. Key point lies in using different threshold for different modes obtained by applying EWT. Proposed algorithm has been tested on self-acquired (on BIOPAC®MP150) ECG signals of 20 subjects, each of 12 minutes duration and 360,000 samples with sampling rate 500 Hz and average 31.2 compression ratio (CR) and 3.28% percentage ratio distortion (PRD) have been obtained. Algorithm has also been applied on all 48 arrhythmia signals of Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH)77777\\+766 database, each of 30 minutes duration and 650,000 samples, sampled at 360 Hz and an average 33.1 CR and 3.3% PRD are obtained.

ELECTROCARDIOGRAM BEAT CLASSIFICATION USING S-TRANSFORM BASED FEATURE SET

In this paper, the conventional Stockwell transform is effectively used to classify the ECG arrhythmias. The performance of ECG classification mainly depends on feature extraction based on an efficient formation of morphological and temporal features and the design of the classifier. Feature extraction is the important component of designing the system based on pattern recognition since even the best classifier will not perform better if the good features are not selected properly. Here, the S-transform (ST) is used to extract the morphological features which is appended with temporal features. This feature set is independently classified using artificial neural network (NN) and support vector machine (SVM). In this work, five classes of ECG beats (normal, ventricular, supra ventricular, fusion and unknown beats) from Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia database are classified according to AAMI EC57 1998 standard (Association for the Advancement of Medical Instrumentation). Performance is evaluated on several normal and abnormal ECG signals of MIT-BIH arrhythmias database using two classifier techniques: ST with NN classifier (ST-NN) and other proposed ST with SVM classifier (ST-SVM). The proposed method achieves accuracy of 98.47%. The performance of the proposed technique is compared with ST-NN and earlier reported technique.

Detection of R Peak in Electrocardiogram

Electrocardiogram (ECG) is the recording of electrical activity of the heart. Accurate measurement of ECG parameters is an important requirement of ECG analysis, particularly if the results are to be used for clinical purposes. Most automatic ECG diagnosis techniques require an accurate detection of the QRS complex. Almost all the analysis methods use the standard Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) database which is available in the digital form. In this paper a method for detection of R peak of ECG signal is discussed. The main objective of this work is to test this method using ECG of Indian patients. ECG signals from Modified Lead II (MLII) are chosen for processing.

R-Peak Detection using Daubechies Wavelet and ECG Signal Classification using Radial Basis Function Neural Network

This paper employed the Daubechies wavelet transform (WT) for R-peak detection and radial basis function neural network (RBFNN) to classify the electrocardiogram (ECG) signals. Five types of ECG beats: normal beat, paced beat, left bundle branch block (LBBB) beat, right bundle branch block (RBBB) beat and premature ventricular contraction (PVC) were classified. 500 QRS complexes were arbitrarily extracted from 26 records in Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH) arrhythmia database, which are available on Physionet website. Each and every QRS complex was represented by 21 points from p1 to p21 and these QRS complexes of each record were categorized according to types of beats. The system performance was computed using four types of parameter evaluation metrics: sensitivity, positive predictivity, specificity and classification error rate. The experimental result shows that the average values of sensitivity, positive predictivity, specificity and classification error rate are 99.8%, 99.60%, 99.90% and 0.12%, respectively with RBFNN classifier. The overall accuracy achieved for back propagation neural network (BPNN), multilayered perceptron (MLP), support vector machine (SVM) and RBFNN classifiers are 97.2%, 98.8%, 99% and 99.6%, respectively. The accuracy levels and processing time of RBFNN is higher than or comparable with BPNN, MLP and SVM classifiers.

An efficient new method for the detection of QRS in electrocardiogram

A simple and efficient new method for QRS detection in Electrocardiogram (ECG) is proposed in this paper. Initially data is preprocessed using two stage median filter for removing baseline drift. The second stage enhances the peaks of ECG wave components by using sixth power of signal. The next stage identifies the QRS complex by taking a variable window size. The detection sensitivity (Se) and positive predictivity (+P) of CSE (Common Standards for Quantitative Electrocardiography) measurement database, MIT/BIH (Massachusetts Institute of Technology/Beth Israel Hospital) Arrhythmia database, European ST-T database and QT database are Se 99.51 & +P 99.69%, Se 99.21 & +P 99.34%, Se 99.53 & +P 99.72% and Se 99.87 & +P 99.95% respectively. These four standard databases used to perform QRS detection considered 368 cases, tested 1,006,168 beats and achieved overall average sensitivity 99.52% and positive predictivity 99.69%. The MIT/BIH Noise Stress Test Database also tested by proposed method.

Optimization of a multi-class MLP ECG classifier using FCM

ECG pattern classification using MLP is an effective and robust technique. Due to the inherent structure of MLP and training algorithm, MLPs tend to be slow and bulky in terms of the hidden layer neurons. This condition is aggravated further if the input data dimension is large as in the case of ECG. The paper addresses this problem by optimizing the MLP using an additional clustering of the feature extracted data. Considerable reduction in the size of MLP was recorded (max. 67.9%) with a reduction in training time (maximum of 59.81%). The experimentation used benchmark arrhythmia database from Physionet Massachusetts institute of technology–Beth Israel hospital (MIT-BIH). Four feature extraction methodologies were subjected to fuzzy c-means clustering for obtaining optimized MLPs. Ten statistical morphological features were also considered for designing the MLP classifiers. Keywords: ECG, MLP, DCT, DWT, FCM, morphological features.