Amirkoushyar Ziabari

I have been working on Physics-Based Image Reconstruction as well as two projects on medical CT image reconstruction using Model Based Iterative Reconstruction (MBIR) and Deep Learning. In the physics-based image reconstruction project, I am using parameters obtained from atomistic simulations of materials along with the phase-field models as prior information to perform soft segmentation and reconstruction of noisy microscopy images. I am also collaborating on two projects with GE Healthcare to develop 1) an MBIR algorithm with spatially adaptive sinogram weights to improve spatio-temporal resolution of Cardiac CT; and 2) an MBIR algorithm that uses deep learning techniques to produce high quality CT images orders of magnitude faster than the state-of-the-art methods.