Abstract

Spectral imaging requires rapid analysis of spectra associated with each pixel. Least-squares fitting using a transport theory is slow. A rapid algorithm has been developed that uses matrix inversion to solve for the absorption spectra of a tissue using a lookup table for photon pathlength based on numerical simulations. The algorithm uses tissue water content as an internal standard to specify the strength of optical scattering. The uncertainty in tissue water content is usually small compared to the uncertainty in tissue scattering properties, but if water content is expected to vary then another method is needed to document tissue water. When implemented in MATLAB, the method is ~100-fold faster than using fminsearch().

© 2010 Optical Society of America