CS Events

PhD Defense

Integrated Reconstruction, Segmentation and Functional Analysis for MRI Data


Download as iCal file

Thursday, March 18, 2021, 02:00pm - 04:00pm


Speaker: Qiaoying Huang

Location : Remote via Webex


Prof. Dimitris Metaxas (Advisor)

Prof. Konstantinos Michmizos

Prof. Desheng Zhang

Dr. Zhen Qian (Tencent America)

Event Type: PhD Defense

Abstract: Magnetic Resonance Imaging (MRI) plays an important role in many clinical applications due to its low radiation and high-contrast imaging. Traditional sequential pipeline for analyzing MRI data includes three steps: reconstruction, segmentation, and functional analysis. For dynamic MRI, i.e., dynamic cardiac MRI, the motion estimation step is also included. These tasks are highly entangled and share important information. However, current approaches are either computationally inefficient or require lots of human efforts to design complicated neural architectures. Moreover, limited work has explored the relationship between different tasks and utilized the coherence information between them. In this thesis, we propose 1) more effective and efficient deep learning-based MRI reconstruction methods; 2) integrated approaches for joint reconstruction and segmentation as well as joint reconstruction and motion estimation; 3) thickness and thickening estimation methods of the left ventricle wall for the diagnosis and characterization of different cardiac diseases. The proposed methods are extensively validated on four different medical applications: 2D cardiac and brain MRI reconstruction, 2D cardiac MRI reconstruction and segmentation, dynamic 3D cardiac MRI reconstruction and motion estimation, dynamic 3D cardiac MRI thickness and thickening estimation. Our proposed approaches are evaluated on various datasets and have outperformed current state-of-the-art methods in all these studies.


Link: https://rutgers.webex.com/rutgers/j.php?MTID=mfc9d1a7af52679e2fb1b1edc390ded7e

Meeting number: 120 393 8145
Password: h84RU6iQy7Q