Szczegóły publikacji
Opis bibliograficzny
A continuous approach to the ECG noiseprint estimation / Piotr AUGUSTYNIAK // Bio-Algorithms and Med-Systems / Jagiellonian University. Medical College ; ISSN 1895-9091. — 2011 — vol. 7 no. 4, s. 51–56. — Bibliogr. s. 55–56, Abstr.
Autor
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 66221 |
|---|---|
| Data dodania do BaDAP | 2012-06-05 |
| Tekst źródłowy | URL |
| Rok publikacji | 2011 |
| Typ publikacji | artykuł w czasopiśmie |
| Otwarty dostęp |  |
| Creative Commons | |
| Czasopismo/seria | Bio-Algorithms and Med-Systems |
Abstract
Signal quality is a common problem in biomedical applications, as it impacts the reliability of electrocardiogram interpretation. The demand for a need for a dependable signal-to-noise measure is reinforced by widespread telemedical recordings in the home care conditions being interpreted automatically. Present techniques are based on baseline noise measurement and assumption about temporal noise stability. This paper presents an alternative approach to ECG noiseprint estimation technique based on a noise model calculated from seamless time-frequency representation. The key principle is recognition of area for possible cardiac components with use of temporarily adapted local bandwidth variability function. The part free of cardiac influence, above the local bandwidth of the ECG, represents background activities of any origin (muscle, mains interference etc.). Next, we consider the non-uniformly sampled time series in each particular scale as piecewise noise estimate and apply interpolation techniques to estimate the noiseprint in regions containing the ECG components. The algorithm was implemented and tested with use of the CSE Database records with the addition of the MIT-BIH Database noise patterns. The differences between the added and estimated noise show similar performance of baseline-based and noise model-based methods (accuracy of 0.69 dB and 0.64 dB respectively) as long as the noise level is stable.
