Pan-Tompkins Research Articles

A new emotion detection algorithm using extracted features of the different time-series generated from ST intervals Poincaré map

Emotion recognition using biological signals plays an essential role in studying psychological states. In many studies, the distinct effect of each physiological signal usually is ignored. By limiting the physiological signals and just evaluating the ECG signals, the paper aims to study the dynamical behavior of the Poincaré map in the form of the feature extraction from the generated time-series. For this purpose, the detection of emotional states in the two-dimensional model of emotion (Arousal-Valence) is utilized using the electrocardiogram signals recorded on the MAHNOB-HCI tagging database. So after signal processing, the waves of Q, R, S, and T are detected by processing the ECG signals using the Pan-Tompkins algorithm. The Poincaré map is adopted for the RR, QT, and ST intervals, then five different time-series are extracted from this mapping, subsequently, various features are extracted from these different time-series in the time domain, frequency domain, time-frequency domain, and nonlinear domain analysis. Finally, the classification of emotional states is performed using three classifiers: KNN, SVM, and MLP. In this study, the extracted optimal features of the different time-series generated from Poincaré map of the ST Intervals are achieved the best average accuracies of 82.17 % ± 4.73 and 78.07 % ± 3.59 in the arousal and valence model, respectively. The superiority of the obtained results from the extracted features of the five different time-series generated from ST-Intervals Poincaré map expresses that in comparison with RR and QT Intervals, the ST Intervals are more affected by ANS response to the emotional stimuli.

Automated QT Interval Measurement Using Modified Pan-Tompkins Algorithm with Independent Isoelectric Line Approach

The QT interval on the electrocardiogram (ECG) signal is known to have an important role in monitoring heart’s electrical activity because the presence of QT interval prolongation can be associated with life-threatening cardiac events. This interval can be identified and measured using either manual or automated techniques. Currently, studies on automated QT interval measurement algorithms are becoming a growing field, as they can provide the best solution to overcome misdiagnosis and timely issues resulting from manual identification. However, the physiological variability of the QRS complex and the fluctuation of the isoelectric line are prevalent issues that need to be considered in the automatic method. In this report, an algorithm to identify the QRS onset and T-wave offset for measuring the corrected QT interval (QTc) is proposed. This method uses an improved Pan-Tompkins algorithm from the previous work with independent of the isoelectric line approach for detecting the QRS onset and the T offset. The algorithm was implemented in Matlab and applied to the 60 seconds duration of 27 records in the PPUKM database with a sampling frequency of 500 Hz. The performance of the algorithm achieved a sensitivity of 100% for QRS onset detection and 100% for T offset detection. As for the accuracy, the algorithm’s performance obtained 100% for QRS onset detection and 99.56% for T offset detection. The mean error results with respect to manual annotation were 37±18.5 ms for QRS onset detection and 32±22.3 ms for T offset detection which was within ANSI/AAMI-EC57:1998 standard tolerance. The proposed algorithm exhibits reliable automated QTc measurement. Besides insensitive to morphological variations of ECG waves, the computational method is simple and possibly implemented as the basis for future software development for portable device applications.

Patient data hiding into ECG signal using watermarking in transform domain

Electrocardiogram (ECG) watermarking provides secure communication of patient information lies in a 1D- ECG signal. The primary challenge in ECG watermarking is the deterioration of an ECG signal which causes the loss and impotence to extract patient information. This paper proposes a wavelet method based watermarking scheme for patient information hiding in the ECG as a QR image. Here, we first convert the 1D-ECG signal to 2D-ECG image using the Pan–Tompkins algorithm. We use a wavelet transform to decompose 2D-ECG image. Wavelet analysis can capture the subtle underlying information of the ECG. Then we further decompose the detail coefficient of wavelet and the QR image using QR decomposition for embedding data. The embedding factor value calculation is adaptive by harnessing the entropy value of the signal. The hidden data is easily extractable with no distortion at the extractor side. The ECG data we use in this paper is from the MIT-BIH database. The results on this dataset suggest that our proposed approach is useful in patient information data hiding scheme in ECG. The proposed method outperforms the state-of-the-art.

Atrial Fibrillation Identification through ECG Signals

This paper presents an algorithm formulated to identify the atrial fibrillation complications through electrocardiogram (ECG) signals. The ECG data for the study was retrieved from Physio Net which consists of normal, atrial fibrillation and other rhythms. The Discrete Wavelet Transform (DWT) was used to remove baseline wanders. Pan Tompkins algorithm was utilized to detect the P, Q, R, S and T peak and thus the ECG signals were segmented based on each cycle. The morphological features were extracted directly from the time-series while statistical features were extracted after Stockwell transform (S-transform) was applied to the data. Genetic Algorithm (GA) and reliefF algorithm have been applied separately to select the optimum features for classification purpose. Bagged Tree ensemble algorithm, Decision Tree and k-Nearest Neighbour (KNN) algorithm were used as classifiers to identify atrial fibrillation through ECG signals. The classification results with and without feature selection techniques are presented. Prior to the feature selection, Bagged Tree is the classifier best performing classifier with 86.50% of accuracy, 84.38% of sensitivity and 91.94% of specificity. After feature selection, all the three classifiers have almost the same performance which is nearly 100% of accuracy, sensitivity and specificity. This shows that the proposed combinations of algorithms are reliable and able to improve the identification rate of the normal, atrial fibrillation and other rhythms using lesser number of features.

Frequency response masking based FIR filter using approximate multiplier for bio-medical applications

The advancements in medical healthcare networks and bio-medical sensor technologies enabled the use of wearable and body implantable intelligent devices for healthcare monitoring. These battery-operated devices must be capable of very low power operation for ensuring long battery life and also to prevent intense radiations. The major power consuming part of these devices are the multipliers built into the digital filters for performing signal processing operations. This paper proposes a low power signed approximate multiplier architecture for bio-medical signal processing applications. The circuit characteristics and error metrics of the proposed multiplier are estimated to verify its performance advantage over other approximate multipliers. In order to validate the efficacy of the approximate multiplier in real time signal processing applications, a band pass finite impulse response filter (FIR) filter is designed using frequency response masking technique and used in the Pan Tompkins method for the extraction of QRS complex from raw ECG data. The sensitivity, positive predictivity, and detection error rate of the QRS detection method are estimated and the results show that the approximate filtering method implemented gives a comparable performance as that of exact methods.

An Efficient Neural Network Model for the Identification of Stress using Electrocardiogram

People are facing numerous pressures in their daily routine in the latest society. Stress has traditionally has been described as action from a calm state to an emotional state in order to preserve the integrity of organism. Stress observation is very important for mental wellbeing and early identification of stress related disorders. Stress is to learn the body response in stressful state, whenever the body reaction is activated that means the heart rate and blood pressure will raise and several hormones enter our bloodshed. These hormones and bodily changes may increases our performances to a particular extent. Everyone's response to stress is discreet, and not all stress is bad. Someone may discover a significant condition of pressure to be enjoyable, while others may find it stressful. However, individuals also have different stress symptoms. stress area can also recognize using frequency and excitation of a speech signal, Since the biomedical signals are consistently related to central nervous system, therefore physiological parameters are the best way to understand the human emotions. The present work is focused on stress identification from Electrocardiogram using ECG physiologic net database, then entire environment of ECG signal characteristics i.e. mean heart rate variability (HRV), standard deviation of all R-R interval (SDNN), square root mean of the sum of the square difference between R-R interval (RMSSD) and number of consecutive R-R interval variations greater than 50ms (NN50), these features are extracted using Pan-Tompkins algorithm, then it is trained and validated to machine learning using back-propagation algorithm in neural network model. With the help of these features (mean HRV, SDNN, RMSSD and NN50), the study can be analyzed whether a person is under stress or not. Thus how the suggested technique provides the subjective information which helps the doctor to find out whether the person is under stress or not.

A novel IRBF-RVM model for diagnosis of atrial fibrillation

Background and objectiveAtrial fibrillation (AF) is one of the common cardiovascular diseases, and electrocardiography (ECG) is a key indicator for the detection and diagnosis of AF and other heart diseases. In this study, an improved machine learning method is proposed for rapid modeling and accurate diagnosis of AF. MethodsThis paper presents a novel IRBF-RVM model that combines the integrated radial basis function (IRBF) and relevance vector machine (RVM), which is utilized for the diagnosis of AF. The synchronous 12-lead ECG signals are collected from the human body surface so as to fully reflect the electrical activity of the whole heart. RR intervals of the QRS-waves in ECG signals are obtained by means of the classical Pan–Tompkins algorithm. The RR-features extracted from RR intervals are adopted as the diagnostic features for AF patients. In addition, the conventional RBF-RVM model, support vector machine (SVM) and other machine learning methods are also investigated for the diagnosis of AF so as to reflect the advantage of the proposed IRBF-RVM model. The open MIT-BIH arrhythmia database (MITDB) is also used to evaluate the predictive performance of these state-of-the-art methods. ResultsAltogether 1056 AF patients and 904 healthy people are participated in this study and validate the effectiveness of each channel of the 12-lead ECG signals. Experimental results show that the classification rate of IRBF-RVM can reach up to 98.16% by recurring to Channel II of the 12-lead ECG signals. ConclusionsIRBF-RVM absorbs the advantages of IRBF, which makes the kernel parameter of IRBF-RVM have a much larger selectable region than RBF-RVM. In addition, RVM has faster modeling and recognition speed in comparison with SVM. This work lays the foundation for the application of RVM to accurate diagnosis of AF.

Monitoring of HR and BP using DFT and Pan-Tompkins Algorithm

Monitoring of HR and BP using DFT and Pan-Tompkins Algorithm

Electrocardiogram monitoring system based on Android smartphone and microcontroller unit

Disease of the heart is one of the death main cause in the world. Special attention to prevent more deaths from the heart disease is needed; one of them is early diagnose for heart disease with electrocardiogram (ECG). ECG represent electrical activity of the heart. The heart’s electrical activity can be recorded using some electrodes that is placed on certain body’s surface. ECG records can be used to determine abnormalities of the heart. We design a ECG monitoring system based on android smartphone and microcontroller unit. The ECG signal is obtained directly from patient using AD8232 module. Then, ECG signal is processed in microcontroller unit. Furthermore, the ECG signal is sent to Android smartphone and displayed in this device. We analyse the abnormalities of the heart using the heart rate from ECG signal. Heart rate is obtained from RR-interval in the ECG which represent the time interval between two detected R peak. We use modified Pan-Tompkins algorithm to detect the peak of R wave. The device that we design give good performance to detect and analyse abnormalities of the heart using heart rate.

Light‐weight configurable architecture for QRS detection

In this study, the authors present a configurable architecture having gate count of ≃ 3.2 k and on the fly reconfigurability for low-power biomedical applications such as QRS detection, ExG processing etc. The proposed architecture is a light-weight co-processor that supports on-node digital signal and image processing functions potentially eliminating the power consumed by radios in wireless sensor node and body sensor network. The architecture consists of a 3 × 3 array of register units along with adaptive memory with configurable data path. The architecture can be configured on-the-fly for seven functions with the current memory structure. However, more number of functions can be targeted with increased memory. They demonstrate the realisation of Pan–Tompkins algorithm commonly used for QRS detection on the proposed architecture using the reconfigurability. This work offers ≈ 4 × reduced area and 2.3 × increase in performance with respect to the existing contemporary literature.

Model-based computer-aided method for diagnosis of cardiovascular disease using IR thermogram.

The human thermal system maintains a core body temperature of 36.7°C. Any deviation from this causes considerable discomfort. The arteries and veins play a significant role in transferring the temperature to maintain the thermal comfort of the human system. Thus the aim of this work was to bring out the potential of thermal imaging in diagnosing the cardiovascular disease in comparison with the standard techniques. The average skin surface temperature (SST (°C)) was compared with the standard bio-signaling method electrocardiograph (ECG) and the ECG-derived heart rate variability (HRV) time domain parameters. Standard biochemical assay and one minute ECG signal were obtained from known CVD men (n=10) and age-matched men (n=10) In each subject, infra-red (IR) thermogram of selected skin area of the body was obtained. From the ECG signal, RR interval was computed using Pan-Tompkins algorithm. HRV time domain variables were also calculated. When SST(°C) variables using static SST (°C) alone were utilized in the CAD model for evaluation of CVD, it was found that, Naive-Bayes classifier with variables ranking and selection by Wilcoxon method gave an accuracy of 90%, and its sensitivity and specificity were found to be 80% and 90% respectively. It was comparable to the obtained CAD model using ECG variables as a standard, whose accuracy was same, and its sensitivity and specificity were found to be 90% and 80% respectively. Also, a better accuracy of 95% (sensitivity-90%, specificity-75%) could be achieved in classifying the normal and CVD subjects when the static average SST (°C) was used along with the HRV time domain variables. Thus the inclusion of IR thermogram in the CAD model along with standard variables improves the accuracy of cardiovascular disease evaluation.

A new technique for ECG signal classification genetic algorithm Wavelet Kernel extreme learning machine

The examination and classification of Electrocardiogram (ECG) records have become particularly significant for diagnosing heart diseases. Machine learning methods are widely used in classifying ECG signals. In this study, Physikalisch-Technische Bundesanstalt Diagnostic ECG Database (PTBDB) from Physionet Database was used to classify ECG signals. Pan-Tompkins algorithm and Discrete Wavelet Transform (DWT) methods were used for extracting critical points such as QRS complex, PR, ST and QT of ECG signal. Afterwards, Traditional Extreme Learning Machine (ELM) was implemented to the ECG signal. Finally, Genetic Algorithm on software Genetic Algorithm Wavelet Kernel Extreme Learning Machine was improved for the determination of the coefficients, which were used in the Wavelet Kernel Extreme Learning Machine algorithm, in which wavelet function was accomplished. In this scope, it was observed that best classification performance values were reached to Acc 95%, Se 100% and Spe 80% with the implementation developed with the Genetic algorithm.

Comparison of Normal and Abnormal Conditions in ECG Using RR Variablity and Spectral Density

Abnormality of the heart is monitored by Electrocardiograph (ECG). The ECG waveform is formed of PQRS pattern. Differentiation of the abnormalities based on the ECG signal is simple algorithm for diagnosis. ECG data of normal, atrial fibrillation and congestive heart failure is obtained from a authorized database. R peak from the QRS complex is detected using Pan-Tompkins algorithm for analysis. Mean RR and heart rate variability parameters are extracted from the QRS complex detected. With these results, the difference in the three ECG signals can be determined. For further detailed comparison, frequency component variation is analysed using power spectral density. Based on density spectrum, the differentiation of normal and abnormal ECG signals can be determined.

Robust QRS detection for HRV estimation from compressively sensed ECG measurements for remote health-monitoring systems

Objective: To present a new compressive sensing (CS)-based method for the acquisition of ECG signals and for robust estimation of heart-rate variability (HRV) parameters from compressively sensed measurements with high compression ratio. Approach: CS is used in the biosensor to compress the ECG signal. Estimation of the locations of QRS segments is carried out by applying two algorithms on the compressed measurements. The first algorithm reconstructs the ECG signal by enforcing a block-sparse structure on the first-order difference of the signal, so the transient QRS segments are significantly emphasized on the first-order difference of the signal. Multiple block-divisions of the signals are carried out with various block lengths, and multiple reconstructed signals are combined to enhance the robustness of the localization of the QRS segments. The second algorithm removes errors in the locations of QRS segments by applying low-pass filtering and morphological operations. Main results: The proposed CS-based method is found to be effective for the reconstruction of ECG signals by enforcing transient QRS structures on the first-order difference of the signal. It is demonstrated to be robust not only to high compression ratio but also to various artefacts present in ECG signals acquired by using on-body wireless sensors. HRV parameters computed by using the QRS locations estimated from the signals reconstructed with a compression ratio as high as 90% are comparable with that computed by using QRS locations estimated by using the Pan–Tompkins algorithm. Significance: The proposed method is useful for the realization of long-term HRV monitoring systems by using CS-based low-power wireless on-body biosensors.

Comparison of outlier heartbeat identification and spectral transformation strategies for deriving heart rate variability indices for drivers at different stages of sleepiness

ABSTRACTObjective: Appropriate preprocessing for detecting and removing outlier heartbeats and spectral transformation is essential for deriving heart rate variability (HRV) indices from cardiac monitoring data with high accuracy. The objective of this study is to evaluate agreement between standard preprocessing methods for cardiac monitoring data used to detect outlier heartbeats and perform spectral transformation, in relation to estimating HRV indices for drivers at different stages of sleepiness.Methods: The study analyzed more than 3,500 5-min driving epochs from 76 drivers on a public motorway in Sweden. Electrocardiography (ECG) data were recorded in 3 studies designed to evaluate the physiological differences between awake and sleepy drivers. The Pan-Tompkins algorithm was used for peak detection of heartbeats from ECG data. Two standard methods were used for identifying outlier heartbeats: (1) percentage change (PC), where outliers were defined as interbeat interval deviating >30% from the mean of the 4 previous intervals, and (2) standard deviation (SD), where outliers were defined as interbeat interval deviating >4 SD from the mean interval duration in the current epoch. Three standard methods were used for spectral transformation, which is needed for deriving HRV indices in the frequency domain; these methods were (1) the Fourier transform; (2) an autoregressive model; and (3) the Lomb-Scargle periodogram. The preprocessing methods were compared quantitatively and by assessing agreement between estimations of 13 common HRV indices using Bland-Altman plots and paired Student's t-tests.Results: The PC method detected more than 4 times as many outliers (0.28%) than SD (0.065%). Most HRV indices derived using different preprocessing methods exhibited significant systematic (P <.05) and substantial random variations.Conclusions: The standard preprocessing methods for HRV data for outlier heartbeat detection and spectral transformation show low levels of agreement. This finding implies that, prior to designing algorithms for detection of sleepy drivers based on HRV analysis, the impact of different preprocessing methods and combinations thereof on driver sleepiness assessment needs to be studied.

Direct Arrhythmia Classification from Compressive ECG Signals in Wearable Health Monitoring System

Due to the capacity of processing signal with low energy consumption, compressive sensing (CS) has been widely used in wearable health monitoring system for arrhythmia classification of electrocardiogram (ECG) signals. However, most existing works focus on compressive sensing reconstruction, in other words, the ECG signals must be reconstructed before use. Hence, these methods have high computational complexity. In this paper, the authors propose a cardiac arrhythmia classification scheme that performs classification task directly in the compressed domain, skipping the reconstruction stage. The proposed scheme first employs the Pan–Tompkins algorithm to preprocess the ECG signals, including denoising and QRS detection, and then compresses the ECG signals by CS to obtain the compressive measurements. The features are extracted directly from these measurements based on principal component analysis (PCA), and are used to classify the ECG signals into different types by the proposed semi-supervised learning algorithm based on support vector machine (SVM). Extensive simulations have been performed to validate the effectiveness of the proposed scheme. Experimental results have shown that the proposed scheme achieves an average accuracy of [Formula: see text] at a sensing rate of 0.7, compared to an accuracy of [Formula: see text] for noncompressive ECG data.

Analysis of the High-Frequency Content in Human QRS Complexes by the Continuous Wavelet Transform: An Automatized Analysis for the Prediction of Sudden Cardiac Death.

Background: Fragmentation and delayed potentials in the QRS signal of patients have been postulated as risk markers for Sudden Cardiac Death (SCD). The analysis of the high-frequency spectral content may be useful for quantification. Methods: Forty-two consecutive patients with prior history of SCD or malignant arrhythmias (patients) where compared with 120 healthy individuals (controls). The QRS complexes were extracted with a modified Pan-Tompkins algorithm and processed with the Continuous Wavelet Transform to analyze the high-frequency content (85–130 Hz). Results: Overall, the power of the high-frequency content was higher in patients compared with controls (170.9 vs. 47.3 103nV2Hz−1; p = 0.007), with a prolonged time to reach the maximal power (68.9 vs. 64.8 ms; p = 0.002). An analysis of the signal intensity (instantaneous average of cumulative power), revealed a distinct function between patients and controls. The total intensity was higher in patients compared with controls (137.1 vs. 39 103nV2Hz−1s−1; p = 0.001) and the time to reach the maximal intensity was also prolonged (88.7 vs. 82.1 ms; p < 0.001). Discussion: The high-frequency content of the QRS complexes was distinct between patients at risk of SCD and healthy controls. The wavelet transform is an efficient tool for spectral analysis of the QRS complexes that may contribute to stratification of risk.

ECG beat classification using machine learning techniques

An arrhythmia is an abnormality in the heart rhythm, or heartbeat pattern. ECG beats can be classified into six different arrhythmia beat types (left bundle branch block, right bundle branch block, paced beats, premature ventricular contradiction, atrial premature beats, and normal rhythm). Early and accurate detection of arrhythmia types is important in detecting heart diseases and choosing appropriate treatment for a patient. This paper proposes a combination of Particle Swarm Optimisation (PSO) and Feed Forward Neural Network (FFNN) for ECG beat classification. We have used MIT-BIH arrhythmia database for data collection and prepared three different datasets. Features, such as R peak sample number and QRS complex, are extracted using Pan-Tompkins algorithm. The extracted features are used as inputs to three different classifiers: Multi-Layer Perceptron Neural Network (MLPNN), Support Vector Machine (SVM), and PSO-FFNN. Results show high classification accuracy of over 97% with either of these three classifiers. The performance comparison of these classifiers is carried out using three measures: sensitivity, specificity, and accuracy. The results suggest that PSO-FFNN shows slightly better performance than MLPNN and SVM in terms of accuracy on all datasets.

A real time ECG signal processing application for arrhythmia detection on portable devices

Arrhythmia describes the disorders of normal heart rate, which, depending on the case, can even be fatal for a patient with severe history of heart disease. The purpose of this work is to develop an application for heart signal visualization, processing and analysis in Android portable devices e.g. Mobile phones, tablets, etc. The application is able to retrieve the signal initially from a file and at a later stage this signal is processed and analysed within the device so that it can be classified according to the features of the arrhythmia. In the processing and analysing stage, different algorithms are included among them the Moving Average and Pan Tompkins algorithm as well as the use of wavelets, in order to extract features and characteristics. At the final stage, testing is performed by simulating our application in real-time records, using the TCP network protocol for communicating the mobile with a simulated signal source. The classification of ECG beat to be processed is performed by neural networks.

Engglis

Electrocardiography (ECG) is a non-invasive technique that is used to diagnose heart abnormalities. ECG records all heart activities and represent them using bio electric signals. Arrhythmia is one of the cardiac disorder that can be detected using ECG. Arrhythmia need to be detected early because of an early symptom of heart disease as deadly as coronary heart disease and heart failure. Arrhythmia described using the difference between the R-peaks based on QRS complex. Therefore, R-peaks detection will be an important factor that can be used to classify arrhythmia disease. One of the widely used methods to detect R-peaks is Pan-Tompkins method. Pan-Tompkins method used a threshold value approach to get all location of R-peaks point from the ECG signals. This study proposed a development based on Pan-Tompkins method by change the threshold value using normalize technique and moving windows approach to get all location of R-peaks point from the ECG signals. This study uses MIT-BIH arrhythmia dataset. This method can show the R-peaks detection with 99.83% sensitivity and 0.40% total error rate detection. Hence, this method has potential to be used for classifying arrhythmia disorder based on the R-peaks point.