Novel QRS detection based on the Adaptive Improved Permutation Entropy. (February 2023)

• Record Type: Journal Article
• Authors: Mansourian, Nastaran; Sarafan, Sadaf; Torkamani-Azar, Farah; Ghirmai, Tadesse; Cao, Hung
• Is Part Of: Biomedical signal processing and control. Volume 80(2023)Part 1
• Publication Date: 2023-02
• Subjects: Adaptive improved permutation entropy -- ECG analysis -- Health monitoring -- Permutation entropy -- QRS detection -- Signal processing; 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.2022.104270
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Detection of the QRS complex is the most important step in analyzing ECG signals for heart monitoring and diagnosis. There have been several QRS-peak detection methods reported in the literature. Most of these methods have low performance under noisy conditions. In this paper, we propose a novel QRS detection algorithm based on a new Permutation Entropy (PE) method that we developed and referred to as the Adaptive Improved Permutation Entropy (AIPE) method. The parameters of the AIPE method are determined based on the specific signal properties. Implementing the AIPE method leads to prominently preserving the QRS complex and eliminating noises of the ECG signal without smoothing the ECG signal. Our simulations show that the proposed QRS detection algorithm is effective and robust under noisy conditions. The algorithm is validated on the MIT-BIH Noise Stress Test Database for various SNR values. In addition, we examined the algorithm's performance under motion noise conditions, mimicking a practical scenario. We used the metrics of sensitivity, positive predictive, and F1 score to evaluate the performance of our algorithm and compare it with several other algorithms explained in the literature. Our investigation shows that the proposed algorithm is efficient and effective. More importantly, it is robust under noisy conditions providing superior performance over other recent and popular QRS detection algorithms, including the popular Pan–Tompkins and the recent Advanced Adaptive Multilevel Thresholding (AAMT) algorithms. Highlights: Use permutation entropy for first time for QRS detection. introduction of two threshold levels dependent to signal to recognize noise and real QRS. Applying motion noise condition to database and test algorithm that shows the suitable performance.