Paper on Electrocardiogram Signal Accepted for Publication

March 10, 2011
School of Engineering, Coimbatore

A paper from Amrita University’s School of Engineering at Coimbatore was recently accepted for publication in Elsevier’s Journal of Biomedical Signal Processing and Control.

Dr. K. P Soman, Professor and Head of the Center for Computational Engineering and Networking (CEN) and Dr. Sabarimalai Manikandan, Assistant Professor at CEN, jointly authored this soon-to-be published research paper titled A Novel Method for Detecting R-peaks in Electrocardiogram (ECG) Signal.

Dr. K. P Soman & Dr. Sabarimalai Manikandan

Cardiovascular disease (CVD) is the most life-threatening disease, and is responsible for half of all deaths in developed and developing countries. Many of these deaths could easily be prevented with early detection and continuous monitoring of cardiac information relating to heart health conditions.

An electrocardiogram (ECG) is a primary diagnostic tool that is used to analyze heart’s electrical activity. A typical ECG signal consists of the P-wave, QRS complexes and T-wave. The features such as amplitudes, durations and shapes of the P-QRS-T-waves and the intervals provide much valuable information for the diagnosis of cardiac diseases.

Automatic detection of the R-peaks in a long-term electrocardiogram (ECG) signal is the most important first step for ECG wave delineation, heart-rate variability (HRV) analysis, heart sound segmentation, medical biometric and ECG coding systems.

The following is an outline of the professors’ new paper.

ecg2Discovery of the R-Peak is pivotal in a large variety of electrocardiogram applications. But almost all existing R-peak detectors suffer from the non-stationarity of both QRS morphology and noise.

To solve this problem, the professors put forth a new R-peak detector which is grounded in the new system of preprocessing and an automated peak-finding chain of reasoning.

The paper presents an unprecedented four-stage methodology for the automated ascertainment of R-peaks in an ECG signal. The preprocessor introduced is based on a bandpass filter, first-order forward derivative, amplitude normalization, Shannon energy estimation, and zero-phase filtering with rectangular impulse response that provides a smooth envelogram of the ECG signal.

The professors first illustrate that the suggested preprocessor containing a Shannon energy envelope estimator is more capable of identifying R-peaks when the following factors are involved: wider- and smaller-QRS complexes, negative QRS polarities and abrupt alterations in QRS amplitudes.

After this illustration, they use the Hilbert-transform and moving average filter to defend the reasonableness of the simplicity and strength of the suggested peak-finding logic.

The effectiveness of the R-peak detector is confirmed by means of the first-channel of the 48 ECG records of the MIT-BITH arrhythmia database. The R-peak has a detection accuracy of 99.80%, sensitivity of 99.93% and positive predictivity of 99.86%. The results of different experiments indicate that the proposed R-peak detection method is considerably more effective than other recognized methods in dealing with noisy or pathological signals. The proposed peak finding task can be used for identifying locations of P/T wave peaks.

R-pekThe proposed approach does not require additional decision rules with sets of thresholds based on the running estimates of the signal peaks and noise peaks, the average RR interval and rate limits, a set of tactics for blanking and T-wave discrimination and training phase.

Reviewers appreciated that the paper was excellent, very well written and interesting to read. It also provides a good review of the topic. One reviewer commented, “First of all, I wish to congratulate the authors for this paper. I have reviewed a lot of papers and it is very rare to read at a first review step a paper so well written and well organized.”

“It has great potential within the field of PCG signal processing for the development of robust, accurate, automated computer-aided diagnostic systems, biometric systems with high recognition accuracy and efficient sound enhancement and coding systems,” explained the article The Heart of the Matter which was published in August 2010 in the biweekly scientific journal Electronics Letters. “The reliable and accurate detection of endpoints of all heart sounds has remained as a very challenging problem,” read the article.

Share this Story: