Diffuse optical techniques have been extensively employed during the last years to retrieve the optical properties of tissue with and without inclusions. The usual approach is to use the diffusion approximation to the radiative transfer equation. However, if low- or non-diffusive regions are inside the studied volume, the diffusion approximation does not hold and the radiative transfer equation needs to be solved, which is computationally much more demanding. In this contribution, the problem of determining the optical properties of the whole volume of a turbid host medium containing both diffusive and non-diffusive inhomegeneities is examined. The situation reproduces clinical cases in which tumors and cysts can be present inside a studied tissue. To achieve this, an extended Kalman filter with compensation by Bayesian error modeling approach was adopted. Applying this technique, the diffusion approximation is used for calculations, reducing computation time, and discrepancies in the non-diffusive regions are compensated by the radiative transfer equation, thus keeping accuracy over the whole volume. The proposal is validated by phantom experiments showing very good results.
Image Reconstruction for Diffuse Optical Tomography Based on Radiative Transfer Equation
