P- and T-wave delineation in ECG signals using parametric mixture Gaussian and dynamic programming. (May 2019)

• Record Type: Journal Article
• Authors: Rao MV, Achuth; Gupta, Prakhar; Ghosh, Prasanta Kumar
• Is Part Of: Biomedical signal processing and control. Volume 51(2019)
• Page Start: 328
• Page End: 337
• Publication Date: 2019-05
• Subjects: ECG -- PT wave -- Dynamic programming; 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.2019.03.001
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Detection and tracking of the P- and T-waves using two mixture Gaussian function and the Dynamic programming are proposed. A key feature of the proposed algorithm is that it allows to incorporate the prior knowledge about the P/T wave location variations and robustness to errors in QRS detection. The proposed algorithm is able to accurately delineate P/T-waves in case of T-wave appears in the middle of a non-QRS region and the U-wave is dominant compared to the T-wave. Abstract: Detection and tracking of the P- and T-waves are important issues in the analysis and interpretation of the ECG signals. This paper addresses the problem by using two mixture Gaussian function and the Dynamic programming. A key feature of the proposed algorithm is that it allows to incorporate the prior knowledge about the P/T wave location variations and robustness to errors in QRS detection. The proposed algorithm is evaluated on the annotated QT-database and compared against the algorithms based on differential evolution optimization strategy (DEOS) and generating blocks of interest (GBI). The experiments show that the proposed method determines the P- and T-peak locations with a root mean square error of 0.085 s and 0.091 s respectively. Both these values are better than the corresponding values from DEOS and GBI. Similarly, the proposed algorithm achieves a sensitivity of 96.13% and predictivity of 97.70%. While the predictivity is higher than both DEOS and GBI, the sensitivity is on par with GBOI and higher than that of DEOS.