Globally Optimal Surface Segmentation in Medical Imaging: From Graph Theory to Deep Learning
Professor Xiaodong Wu
3:55 p.m.–5:10 p.m., October 8, 2020 | Virtual Seminar
To usher in a new era of Precision Medicine, imaging is playing an increasingly significant role, with its superior capability of phenotyping the physical manifestations of diseases.
Meanwhile, this opportunity poses a great challenge — enhancing image processing technology to meet new demands. Robust, efficient, and accurate automated image segmentation is an indispensable step towards quantitative image analysis.
In this talk, Professor Xiaodong Wu will present effective image segmentation methods to optimally identify the boundary surfaces of the target objects, based on enabling graph algorithms and emerging deep learning.
His team’s model-based deep-learning (MoDL) framework for image segmentation unifies the strengths of both deep learning and the graph-based surface segmentation model in a single framework while minimizing their respective limitations.
The novel feature of their method is designed with a focus on the global optimality of the segmentation solutions. Examples and segmentation results on various medical image datasets will be shown.
Professor Wu received BS and MS degrees both in Computer Science from Peking University, China, in 1992 and 1995, respectively. He earned his Ph.D. degree in Computer Science and Engineering from the University of Notre Dame in 2002.
He is currently a professor in the Departments of Electrical and Computer Engineering and Radiation Oncology at the University of Iowa. His research interests are primarily in the areas of computational precision medicine, biomedical imaging, and computer algorithms. He has published over 120 papers in those areas and holds several U.S. patents on technical inventions for radiation cancer treatment and biomedical imaging.
Professor Wu is a recipient of the prestigious NSF CAREER Award (2009) and the NIH K25 Career Development Award (2007). His research is currently under the support of both NIH and NSF.