A study on numerical simulation of image reconstruction in optical computer tomography

Abstract

Time-resolved optical computer tomographic imaging has received considerable interest recently because of its non-invasiveness and non-destructiveness to biological tissue, and several attempts have been made for its implementation. The image recovery algorithm described in this paper is based on the finite-element method solution to the diffusion equation as the photon transport model. By formulating image reconstruction as an optimization problem with the objective function expressed by the error norm between actual and modeled measurement sets, this algorithm incorporates a direct search optimization method, the rotating coordinates method, into the solution strategy to avoid excessive computation of the time-consuming forward problem. Several numerically simulated results for the images of the absorption and scattering coefficient distribution within a circular tissue, reconstructed from the integrated intensity, the mean time of photon flight or their weighted sum, are given. The influence of the numbers of stimulating sources and boundary detection points on the resulting images is discussed. Finally, this paper concludes that the reconstruction algorithm used is feasible but needs to be improved in its accuracy and efficiency.