[Wavelength selection of the oximetry based on test analysis of variance].

In order to improve the precision and reliability of the spectral measurement of blood oxygen saturation, and enhance the validity of the measurement, the method of test analysis of variance was employed. Preferred wavelength combination was selected by the analysis of the distribution of the coefficient of oximetry at different wavelength combinations and rational use of statistical theory. Calculated by different combinations of wavelengths (660 and 940 nm, 660 and 805 nm and 805 and 940 nm) through the clinical data under different oxygen saturation, the single factor test analysis of variance model of the oxygen saturation coefficient was established, the relative preferabe wavelength combination can be selected by comparative analysis of different combinations of wavelengths from the photoelectric volume pulse to provide a reliable intermediate data for further modeling. The experiment results showed that the wavelength combination of 660 and 805 nm responded more significantly to the changes in blood oxygen saturation and the introduced noise and method error were relatively smaller of this combination than other wavelength combination, which could improve the measurement accuracy of oximetry. The study applied the test variance analysis to the selection of wavelength combination in the blood oxygen result measurement, and the result was significant. The study provided a new idea for the blood oxygen measurements and other related spectroscopy quantitative analysis. The method of test analysis of variance can help extract the valid information which represents the measured values from the spectrum.

[Application of EMD algorithm to the dynamic spectrum non-invasive measurement of hemoglobin].

Empirical mode decomposition (EMD) algorithm combined with the theory of dynamic spectrum extraction at frequency domain was applied to the noninvasive measurement of hemoglobin concentration. Fifty seven cases' photoplethysmography was collected in the range of 636.98-1086.86 nm in vivo. After the denoising preprocess through the EMD method for each wavelength pulse wave of each sample separately, dynamic spectrum of each sample was made up of all peaks extracted by Fourier transform. Partial least squares regression model was used to establish the calibration and prediction of hemoglobin concentration. Compared to the modeling results without EMD, the correlation coefficient of predicted values and the real values was increased from 0.8798 up to 0.9176. The root mean square error of prediction set was reduced from 6.6759 to 5.3001 g x L(-1) and the relative error was reduced from 8.45 to 6.71%. The modeling accuracy has been greatly improved. The results showed that EMD algorithm can be effectively applied to denoise the spectral data and improve the accuracy of the non-invasive measurement of blood components.

[Increasing SNR of dynamic spectrum method using harmonic waves].

The sources of noise in dynamic spectrum (DS) method and their corresponding frequency characteristic were analyzed in order to improve the signal to noise ratio (SNR) of DS method. The processes of DS data in frequency domain were reviewed by means of energy, then the harmonic waves of DS data were taken into account in the DS signal and some experiments were done to test whether the modified method works well. In addition, corresponding experiments were carried out to seek the relationship between the SNR and the number of harmonic waves, and to determine how many harmonic waves should be involved in order to get the best SNR in DS method. Results showed when harmonic waves were used in the method properly, the modified method can distinguish DS more precisely. And it was also showed that as the number of harmonic waves increased, the correlation coefficient of DS data from different volunteers became smaller at first and then bigger later, while the correlation coefficient of DS data from different sampling site of the same volunteer kept increasing all the time. When the number of harmonic waves was set to 5, the correlation coefficient of DS data from different volunteers goes from 0.73752 to 0.73676, while the one from the same volunteer goes from 0.99416 to 0.99533, which means the modified method can reflect the information about blood component more precisely than the old ones, and thus the SNR of DS reaches the highest.

[Compensation of backscattering coefficient in Fourier-domain optical coherence tomography].

According to the phenomenon that the intensity of backscattering light decreases dramatically as the detecting depth increases, causing blurred imaging in Fourier-domain optical coherence tomography (FDOCT), a compensation method used for different detecting depth was advanced in the present paper. Through compensating the incident power and scattering coefficient of back-scattered light in each layer of the sample, the real gray level of the image was obtained. With the model of multi-layered homogeneous tissue, the main factors influencing image gray were analyzed. Establishing the FDOCT system with multi-layer cover-glass to be the sample, the structure image was obtained. On the basis of cover-glass as the sample, the general formulas for multi-layer compensation were derived. Combining the theoretical compensation and experimental situation, the obtained data were compensated from incident and emergent intensity. The results show that this compensation method is an effective and feasible way to make the deep layer more visible.