Reconstruction of Tissue Heterogeneity by Near Infrared Imaging: a Novel Algorithm and Benchtop Validation

A novel algorithm has been developed for near infrared (NIR) imaging of tissue optical properties with embedded heterogeneity. The algorithm was based on the optical measurements of the absorption perturbation by a matrix of multiple source detector pairs. Direct superimposition algorithm was used to calculate the two dimensional projected image of the tissue absorption. This absorption map is then compared with that of the homogenous phantom in order to reconstruct the absorption perturbation caused by the embedded object. Benchtop setup has been developed to validate this reconstruction scheme. Simulating tumors made of gelatin cylinders of high absorption coefficients were placed inside a homogenous intralipid phantom of specific background absorption. The depth of the tumor was adjusted at different levels and the diffuse reflectance was measured by a tissue imager consisting of a matrix of 4 sources and 4 detectors. The measured absorption coefficients were compared with the actual tumor absorption coefficients at different tumor depths in order to determine the reconstruction accuracy