Qualitative and quantitative improvement of optical tomographic reconstructed images via spatial deconvolution: three-dimensional case

In diffuse optical tomography (DOT) applications, reconstructed images frequently exhibit undesirable features such as blurring and location bias. At least two basic factors underlie these phenomena. First, there must be an inherent limit on the image quality that is ultimately achievable [1]. Second, they may be consequences of steps that are taken to counter the effects of noise in detector data (e.g., regularization) or to ensure stability and/or accelerate convergence of the inverse problem (e.g., matrix scaling). Distortion of spatial information and low spatial resolution thus can arise as difficult-to-avoid tradeoffs from operations that allow one to solve the inverse problem at all. While for an