Methodology to optimize detector geometry in fluorescence tomography of tissue using the minimized curvature of the summed diffuse sensitivity projections.

The dependence of the sensitivity function in fluorescence tomography on the geometry of the excitation source and detection locations can severely influence an imaging system's ability to recover fluorescent distributions. Here a methodology for choosing imaging configuration based on the uniformity of the sensitivity function is presented. The uniformity of detection sensitivity is correlated with reconstruction accuracy in silico, and reconstructions in a murine head model show that a detector configuration optimized using Nelder-Mead minimization improves recovery over uniformly sampled tomography.