Assessment of Diffuse Optical Tomography Image Reconstruction Methods Using a Photon Transport Model
Imaging of tissue with near-infrared diffuse optical tomography is emerging as a practical method to map hemoglobin concentrations within tissue for breast cancer detection and diagnosis. The accurate recovery of images by using numerical modeling requires an effective image reconstruction method. We illustrate a comparison between two widely used reconstruction methods using finite element modeling in COMSOL Multiphysics® software for 3D forward data generation. The first reconstruction method was born approximation, which was developed in our laboratory. The second method was an image reconstruction package "NIRFAST" that was developed at Dartmouth College. The image reconstruction assessments were performed based on various tumor sizes, depths, and absorption coefficients. Simulation studies demonstrated the capability of each reconstruction method to recover the true values of absorption coefficients.