Fiber-optic evanescent-wave spectroscopy and neural networks: application to chemical blood analysis

Appl Opt. 2000 Jul 1;39(19):3357-60. doi: 10.1364/ao.39.003357.

Abstract

An infrared spectral analysis of human blood serum has been carried out with Fourier transform infrared fiber-optic evanescent-wave spectroscopy. The measured spectra were analyzed by neural network analysis to predict concentrations of cholesterol, creatinine, total protein, urea, and uric acid. The predicted concentrations were compared with results from standard chemical analysis of blood and a good correlation was observed.