A Diffusion–transport Hybrid Method for Accelerating Optical Tomography

It is well acknowledged that the equation of radiative transfer (ERT) provides an accurate prediction of light propagation in biological tissues, while the diffusion approximation (DA) is of limited accuracy for the transport regime. However, ERT-based reconstruction codes require much longer computation times as compared to DA-based reconstruction codes. We introduce here a computationally efficient algorithm, called a diffusion–transport hybrid solver, that makes use of the DAor low-order ERT-based inverse solution as an initial guess for the full ERT-based reconstruction solution. To evaluate the performance of this hybrid method, we present extensive studies involving numerical tissue phantoms and experimental data. As a result, we show that the hybrid method reduces the reconstruction time by a factor of up to 23, depending on the physical character of the problem.