A multiresolution time-dependent entropy method for QRS complex detection. (February 2016)

• Record Type: Journal Article
• Authors: Farashi, Sajjad
• Is Part Of: Biomedical signal processing and control. Volume 24:(2015)
• Page Start: 63
• Page End: 71
• Publication Date: 2016-02
• Subjects: Signal processing -- Event detection -- QRS complex -- Time dependent entropy; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2015.09.008
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: An algorithm based on time dependent entropy is proposed for QRS complex detection. Entropy of ECG signal is calculated with different temporal resolutions. Various QRS morphologies can be detected successfully. The proposed algorithm outperforms traditional algorithms in some aspects like computational complexity and real-time feasibility. Abstract: Electrocardiography is considered as a useful diagnostic tool for evaluating the condition of heart's health. QRS complex, which is produced by depolarization of the heart ventricles, is the main graphical deflection seen on a typical electrocardiogram tracing. Detection of the QRS complexes is the first step toward analyzing the electrocardiogram signal. In this regard, many different algorithms have been proposed so far. In the present work an algorithm based on entropy measure is proposed which uses the calculation of the time dependent entropy for QRS complex detection. The algorithm is implemented in a way that entropy of the electrocardiogram can be calculated in different temporal resolution to improve the accurate detection rate of different QRS morphologies. The MIT-BIH arrhythmia and CSE databases are selected to test the performance of the proposed algorithm. The precision and sensitivity of the proposed method for MIT-BIH database are 99.85% and 99.75%, respectively. Also the detection rate of 99.82% is achieved for CSE database. Additionally, the proposed algorithm is fast enough to be applied in real-time.