Noise Removal in Single-Lead Capacitive ECG With Adaptive Filtering and Singular Value Decomposition

• Published in: IEEE access Vol. 12; pp. 152777 - 152785
• Main Authors: Wang, Kaichen, Tu, Cathy, Man Qian, Chi, Lin, Runwei, Doyle, David, Fujiwara, Yujiro, Wang, Junhong
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adaptive algorithms; Adaptive filtering; Adaptive filters; Biomedical measurement; Biomedical monitoring; capacitive ECG; Correlation coefficients; Electrocardiography; Electrodes; Feasibility studies; Heart rate; Matrix decomposition; morphological analysis; Morphological operations; Noise measurement; Noise monitoring; Noise reduction; Noise sensitivity; Parameter identification; Performance measurement; Recording; Singular value decomposition
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3478779

Capacitive electrodes have been proposed to measure the electrocardiography (ECG) without direct skin contact. Unfortunately, capacitive ECG (cECG) measurements are very sensitive to various types of noise, and are therefore limited for heart-rate monitoring. The aim of the present study is to explore the feasibility of a cECG system with dedicated denoising algorithms for clinical diagnosis. We conducted real cECG measurements on 12 healthy subjects while simultaneously obtaining traditional wet ECG measurements as references. An adaptive canceller and an innovative singular value decomposition (SVD) algorithm were applied to single-lead cECG for noise reduction. The correlation coefficients (CCs) between the processed cECG and the wet ECG were computed with and without the QRS complex as performance metrics. In addition, several clinical parameters were annotated from the detailed waves of both ECG signals by a medical expert and then compared. The results demonstrate high CCs between the cECG and wet ECG for both conditions. Furthermore, the clinical parameters identified from both signals are quite similar. These findings indicate the potential of cECG for morphological analysis in clinical settings.