Dissertation Title: Data-Driven Approaches to Image Reconstruction for X-Ray Computed Tomography
Date: 2023/08/17 - 2023/08/17
Dissertation Title: Data-Driven Approaches to Image Reconstruction for X-Ray Computed Tomography
Speaker: Zhipeng Li, Ph.D. candidate at UM-SJTU Joint Institute
Time: August 17th from 9:00 a.m., 2023 (Beijing Time)
Location: Room 403, Longbin Building
Abstract
X-ray computed tomography (CT) is a popular imaging modality in clinics, and technical advances in CT such as low-dose CT (LDCT) and dual-energy CT (DECT) open up new applications for this imaging modality. This dissertation investigates incorporating data-driven techniques into the model-based image reconstruction (MBIR) and decomposition (MBID) frameworks for LDCT and DECT, respectively. First, an MBID method using unsupervised data-driven priors is proposed for image-domain material decomposition in DECT. Second, an MBID method using supervised data-driven priors is proposed to further improve decompositions. Third, a deep learning-based MBIR method is proposed to obtain high-quality reconstructed images in LDCT.