학술논문

Cuff-Less Blood Pressure Estimation Using Only the ECG Signal in Frequency Domain
Document Type
Conference
Source
2018 8th International Conference on Computer and Knowledge Engineering (ICCKE) Computer and Knowledge Engineering (ICCKE), 2018 8th International Conference on. :147-152 Oct, 2018
Subject
Bioengineering
Communication, Networking and Broadcast Technologies
Computing and Processing
Signal Processing and Analysis
Electrocardiography
Estimation
Blood pressure
Feature extraction
Machine learning algorithms
Frequency-domain analysis
Regression tree analysis
Electrocardiograph
Cuff-less
Whole-based
Mechano- Electric Coupling (MEC)
Language
Abstract
The relationship between the heart's electrical and mechanical activities is expressed by Mechano-Electric Coupling (MEC) term, and blood pressure (BP) is the output of these activities. Electrocardiogram (ECG) is a representation of heart's electrical activity. Previous studies show that there is a nonlinear relationship between the ECG signal and BP values. This paper presents a new algorithm for estimating BP using only the ECG signal and without any cuff. A new feature vector extraction method in the frequency domain, called the frequency whole-based method, is proposed. According to the British Hypertension Society (BHS) standards, the proposed algorithm achieves grade B for the Diastolic Blood Pressure (DBP) and grade C for Mean Atrial Blood Pressure (MAP) estimations on MIMIC II dataset.