Use of Genetic Algorithms to Optimize Fiber Optic Probe Design for the Extraction of Tissue Optical Properties

This paper outlines a framework by which the optimal illumination/collection geometry can be identified for a particular biomedical application. In this paper, this framework was used to identify the optimal probe geometry for the accurate determination of tissue optical properties representative of that in the ultraviolet-visible (UV-VIS) spectral range. An optimal probe geometry was identified which consisted of a single illumination and two collection fibers, one of which is insensitive to changes in scattering properties, and the other is insensitive to changes in the attenuation coefficient. Using this probe geometry in conjunction with a neural network algorithm, the optical properties could be extracted with root-mean-square errors of 0.30 cm^-1for the reduced scattering coefficient (tested range of 3-40 cm^-1), and 0.41 cm^-1 for the absorption coefficient (tested range of 0-80 cm^-1).