Browsing by Author "Acar, Burak"

Now showing 1 - 5 of 5

Open Access
Assessment of information redundancy in ECG signals
(IEEE, 1997-09) Acar, Burak; Özçakır, Lütfü; Köymen, Hayrettin
In this paper, the morphological information redundancy in standard 12 lead ECG channels is studied. Study is based on decomposing the ECG channels into orthogonal channels by an SVD based algorithm and then reconstructing them. Then 7 of 8 independently recorded ECG channels are decomposed and the missing channel is reconstructed from these orthogonal channels. Thus the unique morphological information content of each ECG channel is assessed through the loss of clinical information in the reconstructed signal. A comparison of the clinical parameters measured from the reconstructed and original ECG is reported.
Open Access
Concept of T-wave morphology dispersion
(IEEE, 1999) Acar, Burak; Yi, G.; Malik, M.
The detection of ventricular repolarization abnormalities is widely being done using the QT interval measurements. However, there are both technical and theoretical problems with QT measurements. We propose two robust methods for the quantification of the ventricular repolarization abnormalities: i) The quantification of the inter-lead morphology differences of the T wave (T Wave Morphology Dispersion - TMD) ii) The analysis of the T wave wavefront direction with respect to the QRS complex (Total Cosine R_To_T - TCRT). Sensitivity and specificity of 82% (84%) in supine position and 77% (79%) in standing position were achieved for TMD (TCRT). Both parameters were more reproducible than conventional QT interval based parameters.
Open Access
New techniques for ventricular repolarization and heart rate variability analyses
(2000) Acar, Burak
This thesis is composed of two parts: i) Development of a fully automatic Heart Rate Variability (HRV) analysis method, and ii) development of new methods for ventricular repolarization (T wave) analysis. The first part of this study deals with fully automatic measurement of heart rate variability (HRV) in short term electrocardiograms. In short, HRV analysis is the spectral analysis of the heart rate signal. Presently, all existing HRV analysis programs require user intervention for ectopic beat identification which is essential for reliable HRV analysis. This makes HRV studies in large populations problematic. A fully automatic algorithm to discriminate ventricular and supra-ventricular ectopic beats from normal beats is pre.sented. The method incorporates several approaches and uses three EGG leads. It uses the template matching for the basic morphology check of the QRS complex and the P-wave, the timing information to avoid unnecessary computation and to adjust the thresholds and also looks for a special QRS morphology which is common in ventricular ectopic beats. The method is tested on a set of real ECG recordings and statistically analyzed on the basis of sensitivity and specificity. Its performance using single ECG leads and different triplets of EGG leads is also studied. We have obtained 99% specificity and SVE sensitivity and 98% VE sensitivity and thus concluded that fully automatic HRV analysis is feasible. The second part of this thesis is on ventricular repolarization analysis (T wave analysis). It has been shown that heterogeneity in ventricular repolarization is a mark of abnormality and can be used for risk stratification. Several methods have been proposed to measure this heterogeneity, among which the QT interval measurements are the most popular ones. After a short discussion of the existing methods, we propose three new approaches for T wave analysis, which are aimed to overcome the drawbacks of the existing methods: The spatial and temporal variations in the T wave morphology and the wavefront direction difference between the ventricular depolarization and repolarization waves. All of the descriptors are defined in an ECG decomposition space constructed by Singular Value Decomposition. The spatial variation characterizes the morphology differences between standard leads. The temporal variation measures the change in interlead relations throughout the T wave. The wavefront direction difference quantizes the difference between the progress of the two processes. None of them requires time domain measurements thus avoid the inaccuracies associated with conventional methods. The new methods are compared with the conventional ones in a set of 1100 normal ECGs. The short-term intra-subject reproducibility of the new and the conventional methods is compared in a set of 760 normal (recorded from 76 normal subjects) and 630 abnormal (recorded from 63 HGM patients) EGGs. The new descriptors’ ability to discriminate normal and abnormal EGGs (both in univariate and multivariate models) is also analyzed on the same data set. A two-way blind study conducted on a set of AMI (Acute Myocardial Infarction) patients have shown that the new methods are able to discriminate the high risk group. The conventional methods were shown to be useless in this patient group in a previous study. We have concluded that the new descriptors do not correlate with the conventional ones, are more reproducible, lead to more significant separation between normal and abnormal ECGs in both univariate and multivariate models.
Open Access
Online ECG signal orthogonalization based on singular value decomposition
(1996) Acar, Burak
Electrocardiogram (ECG) is the measurement of potential differences occurring on the body due to the currents that flow on the heart during diastole and systole. Cardiac abnormalities cause uncommon current flows, leading to strange waveform morphologies in the recorded ECG. Since some abnormalities become visible in ECG only during activity, exercise ECG tests are conducted. The sources of noise during an exercise test are electro myogram (EMG) due to increased muscle activity and baseline wander (BW) due to mechanical motion. Frequency band filtering, used to eliminate noise, is not an efficient method for filtering noise because usually frequency spectra of the interference and the ECG overlap. Rather, a fast morphological filter is required. This thesis is focused on an online filtering approach which separates noise and ECG signals without changing the morphology. The redundancy present in standard 12 lead ECG records is made operational by a Singular Value Decomposition based orthogonalization of the input signals. ECG is represented in a minimum dimensional space whose orthogonal complement takes on noise. The signals in this low dimensional space are used to reconstruct the input signals without noise. Noise elimination also improves data compression. A comparative study of the ST analysis of original and reconstructed signals is presented at the end.
Open Access
Supervised machine learning algorithm for arrhythmia analysis
(IEEE, 1997) Güvenir, H. Altay; Acar, Burak; Demiröz, Gülşen; Çekin, A.
A new machine learning algorithm for the diagnosis of cardiac arrhythmia from standard 12 lead ECG recordings is presented. The algorithm is called VFI5 for Voting Feature Intervals. VFI5 is a supervised and inductive learning algorithm for inducing classification knowledge from examples. The input to VFI5 is a training set of records. Each record contains clinical measurements, from ECG signals and some other information such as sex, age, and weight, along with the decision of an expert cardiologist. The knowledge representation is based on a recent technique called Feature Intervals, where a concept is represented by the projections of the training cases on each feature separately. Classification in VFI5 is based on a majority voting among the class predictions made by each feature separately. The comparison of the VFI5 algorithm indicates that it outperforms other standard algorithms such as Naive Bayesian and Nearest Neighbor classifiers.