3D-surface reconstruction method for diffuse optical tomography phantoms and tissues using structured and polarized light

Abstract

In recent years optical methods became increasingly popular for pre-clinical research and small animal imaging. One main field in biomedical optics research is the diffuse optical tomography (DOT). Many new systems were invented for small animal imaging and breast cancer detection. In combination with the progress in the development of optical markers, optical detectors and near infrared light sources, these new systems have become a formidable source of information. Most of the systems detect the transmitted light which passes through an object and one observes the intensity variations on the detector side. The biggest challenge for all diffuse optical tomography systems is the enormous scattering of light in tissues and tissue-like phantoms resulting in loss of image information. Many systems work with contact gels and optical fibers that have direct contact with the object to neglect the light path between surface and detector. Highly developed mathematic models and reconstruction algorithms based on FEM and Monte Carlo simulations describe the light transport inside tissues and determine differences in absorption and scattering coefficients inside.The proposed method allows a more exact description of the orientation of surface elements from semi-transparent objects towards the detector. Using Polarization Difference Imaging (PDI) in combination with structured light 3D-scanning, it is possible to separate information from the surface from that of the subsurface. Thus, the actual surface shape can be determined. Furthermore, overlaying byproducts caused by inter-reflections and multiple scattering can be filtered from the basic image information with this method. To enhance the image quality, the intensity dispersion between surface and camera is calculated and the creation of 3D-FEM-meshes simplified.

Review of structured light in diffuse optical imaging

“…Methods have also been developed to decrease the overall computation time of tomography, such as the initial projection of sinusoidal fringe patterns to construct a priori surface profiles for generation of the forward model 159 – 161 In addition, structured light strategies are well-suited to benefit from CS-based preconditioning of the illumination and detection fields for improved performance 162 . Beyond the potential of using optimal illumination patterns based on theoretical considerations, structured light methodologies are also amenable to implementing strategies to improve the SNR of the measured data set.…”