Dissertation Defense: Shu Yang, "Scatter Based Novel Imaging Systems"

    Thursday, April 28, 2022 - 9:30am

    Join Zoom Meeting


            Meeting ID:864 7984 1090



    ​Imaging is the process of reproduction of an object’s form. Normally, the image of an object is generated by gathering photons from an object pixel directly and concentrate it onto a detector pixel in a 1 to 1 fashion. However, in many special situations in which direct imaging is not available, people are forced to gather scattered photons that are only indirectly related to the target. In this study, we explore two of these special cases and propose two novel scattering-based methods working at different wavelengths. The first method we studied is a novel non-line-of-sight (NLOS) imaging technology. The main stream time-resolved NLOS technologies usually involve emission of photons actively with ultra fast laser. To explore the new front, we proposed a new NLOS imaging system in which only ambient light is used as illumination. Imaging is achieved through ananlysis of scattered photons on the visible wall. We call this new technology "passive NLOS". The second method we studied is a novel X-ray diffraction tomography algorithm. X-ray diffraction tomography gathers off-plane Rayleigh scattered X-ray photons and reconstruct a mapping of material form factor. Previous reconstruction algorithms applied either maximum likelihood estimation or maximum-a-posteriori estimation with spatially smooth prior. In this study, we explored a new reconstruction algorithm that utilized the sparsity in the material form factors. The proposed new algorithm is a modified version of Expectation-Maximization algorithm that works with Group-Tv regularization spatially and  l1-norm regularization spectrally.