Parametric level-set approach for hyperspectral diffuse optical tomography

We explore the development and performance of algorithms for hyperspectral diffuse optical tomography (DOT) in which data from hundreds of wavelengths are collected and used to determine the concentration distribution of chromophores in the medium under investigation. An efficient method is detailed for forming the images using iterative algorithms applied to a linearized Born approximation model. Specifically, a parametric level-set approach is employed to obtain estimates of chromophore concentrations and scattering properties. We report image reconstructions using 126 wavelengths with mean square error around 0.05 for each chromophore in simulated data, a significant improvement over reconstructions using the level-set method and a pixel based method using fewer wavelengths chosen optimally according to the criterion in [1]. We additionally report improvement in reconstruction for phantom measurements where species concentrations are estimated more accurately than images constructed using fewer wavelengths.