High Resolution 3D Image Reconstruction in Laminar Optical Tomography Based on Compressive Sensing

Laminar optical tomography (LOT) combines the advantages of diffuse optical tomography image reconstruction and a microscopy-based setup to allow non-contact imaging at depth up to a few millimeters. However, LOT image reconstruction paradigm is inherently an ill-posed and computationally expensive inverse problem. Herein, we cast the LOT inverse problem in the compressive sensing (CS) framework to exploit the sparsity of the fluorophore yield in the image domain and to address the ill-posedness of the LOT inverse problem. We apply this new approach to thick tissue engineering applications. We demonstrate the enhanced resolution of our method in 3-D numerical simulations of anatomically accurate microvasculature and using real data obtained from phantom experiments. Furthermore, CS is shown to be more robust against the reduction of measurements in comparison to the classic methods for such application. Potential benefits and shortcomings of the CS approach in the context of LOT are discussed.