Open AccessModular continuous wavelet processing of biosignals: extracting heart rate and oxygen saturation from a video signal
Author(s) -
Addison Paul S.
Publication year - 2016
Publication title -
healthcare technology letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.45
H-Index - 19
ISSN - 2053-3713
DOI - 10.1049/htl.2015.0052
Subject(s) - computer science , respiratory rate , continuous wavelet transform , wavelet , speech recognition , heart rate , wavelet transform , artificial intelligence , information retrieval , medicine , discrete wavelet transform , blood pressure , radiology
A novel method of extracting heart rate and oxygen saturation from a video‐based biosignal is described. The method comprises a novel modular continuous wavelet transform approach which includes: performing the transform, undertaking running wavelet archetyping to enhance the pulse information, extraction of the pulse ridge time–frequency information [and thus a heart rate (HR vid ) signal], creation of a wavelet ratio surface, projection of the pulse ridge onto the ratio surface to determine the ratio of ratios from which a saturation trending signal is derived, and calibrating this signal to provide an absolute saturation signal (S vid O 2 ). The method is illustrated through its application to a video photoplethysmogram acquired during a porcine model of acute desaturation. The modular continuous wavelet transform‐based approach is advocated by the author as a powerful methodology to deal with noisy, non‐stationary biosignals in general.