Tampere University of Technology

TUTCRIS Research Portal

A novel generic algorithm for robust physiological signal classification

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Details

Original languageEnglish
Title of host publicationXIV Mediterranean Conference on Medical and Biological Engineering and Computing 2016
Subtitle of host publicationMEDICON 2016, March 31st–April 2nd 2016, Paphos, Cyprus
PublisherSpringer Verlag
Pages1038-1043
Number of pages6
ISBN (Electronic)978-3-319-32703-7
ISBN (Print)978-3-319-32701-3
DOIs
Publication statusPublished - 2016
Publication typeA4 Article in a conference publication
EventMediterranean Conference on Medical and Biological Engineering and Computing -
Duration: 1 Jan 2000 → …

Publication series

NameIFMBE Proceedings
Volume57
ISSN (Print)1680-0737

Conference

ConferenceMediterranean Conference on Medical and Biological Engineering and Computing
Period1/01/00 → …

Abstract

The last decade has witnessed a significant interest in widespread usage of wearable monitoring devices that could provide continuous measurements of physiological parameters. The design and development of these devices has attracted lots of attention in industry and scientific associations. Advanced and miniaturized electronics with signal acquisition technologies provide a possibility for designing only one device for several physiological measurement purposes. Therefore for designing such an automatic system, a simple generic algorithm for physiological signal classification is required. In this paper, a novel generic algorithm for robust physiological signal classification is presented. The architecture of the proposed system includes preprocessing, feature extraction and a neural network method. Our generic algorithm was able to distinguish different physiological signals such as electrocardiogram (ECG), respiratory signal, seismocardiogram (SCG), electromyogram (EMG) and photoplethysmogram with 100% accuracy. The algorithm was also evaluated by noisy signals with 10 and 20 dB levels of added noise and the same results were achieved. The algorithm could be implemented in healthcare monitoring systems and it can provide the possibility of monitoring various physiological signals with only one device.

ASJC Scopus subject areas

Keywords

  • Classifier, Generic algorithm, Neural network, Physiological signals, Wearable devices

Publication forum classification

Field of science, Statistics Finland