Biomedical subjects

Weijian Xu

Publications and source records attributed to Weijian Xu.

3 recordsLinked to original sources

PubMed2005

Processing of pulse oximeter data using discrete wavelet analysis.

A wavelet-based signal processing technique was employed to improve an implantable blood perfusion monitoring system. Data was acquired from both in vitro and in vivo sources: a perfusion model and the proximal jejunum of an adult pig. Results showed that wavelet analysis could isolate perfusion signals from raw, periodic, in vitro data as well as fast Fourier transform (FFT) methods. However, for the quasi-periodic in vivo data segments, wavelet analysis provided more consistent results than the FFT analysis for data segments of 50, 10, and 5 s in length. Wavelet analysis has thus been shown to require less data points for quasi-periodic data than FFT analysis making it a good choice for an indwelling perfusion monitor where power consumption and reaction time are paramount.

Algorithms↗

PubMed2005

An autocorrelation-based time domain analysis technique for monitoring perfusion and oxygenation in transplanted organs.

In designing an implantable sensor for perfusion monitoring of transplant organs the ability of the sensor to gather perfusion information with limited power consumption and in near real time is paramount. The following work was performed to provide a processing method that is able to predict perfusion and oxygenation change within the blood flowing through a transplanted organ. For this application, an autocorrelation-based algorithm was used to reduce the acquisition time required for fast Fourier transform (FFT) analysis while retaining the accuracy inherent to FFT analysis. In order to provide data proving that the developed method is able to predict perfusion as accurately as FFT two experiments were developed isolating both periodic and quasi-periodic cardiac frequencies. It was shown that the autocorrelation-based method was able to perform comparably with FFT (limited to a sampling frequency of 300 Hz) and maintain accuracy down to acquisition times as low as 4 s in length.

Algorithms↗

PubMed2005

Real-time separation of perfusion and oxygenation signals for an implantable sensor using adaptive filtering.

In this paper, an adaptive filtering algorithm to separate signals due to perfusion and oxygenation has been developed using an 810-nm source, in addition to 660-nm and 940-nm sources, as an internal reference due to its limited oxygen sensitivity. The newly developed algorithm was tested using Monte Carlo simulated data to prove the effectiveness of the 810-nm reference and adaptive algorithm. Following the simulation, an in vitro model was developed to test the algorithm that used a blood flow through system wrapped with tissue. The system had the ability to isolate the effects of perfusion and oxygenation and the algorithm accurately captured the changes in these signals with reliable consistency. Using the serosal surface of the swine jejunum, in vivo data was also taken to analyze the algorithms response to fluctuating perfusion levels like that seen in hemorrhaging or failing transplants. The algorithm was able to extract the perfusion information from the oxygenation information in this in vivo study. Overall, it was shown that an adaptive filtering algorithm using an 810-nm reference has provided a means to separate oxygenation and perfusion.

Algorithms↗