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

Motion-tolerant pulse oximetry based on the wavelet transformation and adaptive peak filtering,10.1109/MECBME.2011.5752131,S. Andruschenko,U. Timm,J.

Motion-tolerant pulse oximetry based on the wavelet transformation and adaptive peak filtering  
BibTex | RIS | RefWorks Download
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 con- tinuous 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. ULSE oximetry relates to continuous non-invasive de- termination of the peripheral oxygen saturation (SpO2) of blood and pulse rate, based on the analysis of photo- plethysmographical (PPG) pulses. Accuracy of the pulse oximetric (PO) readings can be affected by rapid desatura- tion, high levels of other than oxy/deoxygented hemoglobin constituents (i.e. carboxyhemoglobin or methemoglobin) and noise artifacts. The most common sources of noise in PO are the ambient light, electromagnetic coupling from other electronic instruments and motion disturbances. Motion ar- tifacts caused by patient movement are usually the most difficult type of noise to eliminate since their morphology often resembles the PPG-waves and could overlap a usefull signal resulting in low signal-to-noise ratio (SNR) in case of strong disturbances. During persistent severe motion the pulse waveform could be so corrupted, that the pulsative component is no longer possible to identify by conventional methods. This research is being conducted within a framework of the project PHOTOSENS and is funded by the Ministry of Economics, Labour and Tourism of the state Mecklenburg-Western Pomerania, Germany.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.