A Survey on Deep Learning-based Medical Image Registration

PhD Qualifying Examination


Title: "A Survey on Deep Learning-based Medical Image Registration"

by

Miss Ziyi HE



Abstract:

Image registration is the process of estimating the optimal spatial 
transformation between images with matched content. It is a fundamental 
and essential task in medical image analysis and clinical studies like 
segmenting images to find lesions, fusing images from different modalities 
or subjects, quantifying the change of organs during the treatment, and so 
on. Traditional optimization-based registration methods are time-consuming 
and have the risk of falling into local minima. In recent years, the boom 
in deep learning has brought breakthroughs to image registration and many 
other fields in computer vision.

In this survey, we will introduce state-of-the-art learning-based image 
registration methods including both pairwise and groupwise registration. 
Furthermore, as complementary tasks share mutual information that benefits 
each other's learning, we also review some joint frameworks that involve 
registration and other image processing tasks like segmentation. In the 
end, we summarize and discuss future trends in learning- based medical 
image registration approaches.


Date:			Tuesday, 3 November 2020

Time:                  	2:00pm - 4:00pm

Zoom meeting:           https://hkust.zoom.us/j/2769044281

Committee Members:	Prof. Albert Chung (Supervisor)
 			Prof. Pedro Sander (Chairperson)
 			Dr. Qifeng Chen
 			Dr. Xiaojuan Ma


**** ALL are Welcome ****