Motion-tolerant pulse oximetry based on the wavelet transformation and adaptive peak filtering

S. Andruschenko, U. Timm, J. Kraitl, E. Lewis, H. Ewald

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

Abstract

A novel motion-tolerant algorithm for continuous real-time monitoring of blood constituents by means of pulse oximetry is introduced. Motion artifacts frequently lead to false interpretations of the measured signal or can cause a failure of the signal detection. Therefore these disturbances are required to be recognized and suppressed while the useful signal should remain possibly unaffected. The technique is based on the continuous wavelet analysis combined with optional adaptive peak filtering to optimally estimate the physiological parameters. Presented algorithm appears to be a sensitive nonlinear method of processing the pulsative arrhythmic patterns in frequency domain. Reconstruction of the motion-corrupted PPG-waveform could allow an elicitation of the individual clinical parameters which yield additional data about the human health status. The method is not limited to non-invasive oximetry only and can be utilized in other medical fields of patient monitoring.

Original languageEnglish
Title of host publication2011 1st Middle East Conference on Biomedical Engineering, MECBME 2011
Pages324-327
Number of pages4
DOIs
Publication statusPublished - 2011
Event2011 1st Middle East Conference on Biomedical Engineering, MECBME 2011 - Sharjah, United Arab Emirates
Duration: 21 Feb 201124 Feb 2011

Publication series

Name2011 1st Middle East Conference on Biomedical Engineering, MECBME 2011

Conference

Conference2011 1st Middle East Conference on Biomedical Engineering, MECBME 2011
Country/TerritoryUnited Arab Emirates
CitySharjah
Period21/02/1124/02/11

Keywords

  • frequency domain
  • motion-tolerant
  • pulse oximetry
  • real-time
  • wavelet analysis

Fingerprint

Dive into the research topics of 'Motion-tolerant pulse oximetry based on the wavelet transformation and adaptive peak filtering'. Together they form a unique fingerprint.

Cite this