Wavelet Sub-bands Features-based ECG Signal Quality Assessment Scheme for Computer-aided Monitoring System

Abstract

Electrocardiogram (ECG) signal quality assessment (SQA) is a crucial step for improving the performance and reliability of medical diagnostics in computer-aided monitoring systems. Most of the existing ECG SQA methods utilize QRS complex and R-peak information-based features, where the quality assessment performance degrades due to improper detection of R-peaks. Unlike the existing machine learning-based ECG SQA approaches, this work proposes an efficient wavelet sub-bands features-based SQA scheme by avoiding the utilization of R-peak-based features. Sub-band levels features are extracted by incorporating the frequency localization property of wavelet transform (WT). Furthermore, the features are applied to machine learning-based classifier to assess the signal quality. Simulation study and results show that the proposed scheme exhibits an assessment accuracy of 98.11% on the simulated dataset, 99.12% on single channel ECG dataset of Physionet: Computing in Cardiology Challenge (PCinCC) and 97.32% on real arrhythmia dataset: Massachusetts Institute of Technology-Beth Israel Hospital database (MITDB). An in-depth study of the results illustrates that the proposed ECG SQA technique can be utilized for improving the reliability of modern automated computer-aided cardiac monitoring systems.

KEYWORDS:

• Electrocardiogram
• Random forest
• Signal quality assessment
• Sub-bands features
• Wavelet

Acknowledgements

The authors gratefully acknowledge the funding for this research from Ministry of Electronics and Information Technology (MeitY), Govt. of India under the grant number PhD-MLA/4(13)/2015-16.

Additional information

Funding

This work was supported by Ministry of Electronics and Information technology: [Grant Number PhD-MLA/4(13)/2015-16.].

Notes on contributors

Manas Rakshit

Manas Rakshit received his BTech degree from University of Kalyani, West Bengal. He obtained his MTech and PhD both from NIT, Rourkela, Odisha, India. He is currently working as faculty in Department of Electronics and Telecommunication Engineering in IIEST Shibpur India. His research interests include Bio-medical signal processing, machine learning.

Susmita Das

Susmita Das received her BSc Engineering from CET, Bhubaneswar, Odisha, and MSc Engg and PhD both from NIT, Rourkela, Odisha, India. She is currently professor in Department of Electrical Engineering in NIT, Rourkela, India. She is a life member of ISTE and Institute of Engineers, India. She is also a member of IETE, India and IEEE. Her research interests include signal processing and wireless communication. Email: [email protected]