Deep Learning-based Multimodal Medical Image Registration: A Survey

PhD Qualifying Examination


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

by

Mr. Chi Wing MOK


Abstract:

Aligning two or multiple medical images with different modalities through 
robust image registration is crucial to many clinical tasks such as brain 
atlas creation, pre-surgical plan, and in-surgical tumor localization. 
Unlike unimodal registration, robust multimodal registration remains a 
challenging problem in the medical industry due to the non-linear spatial 
correspondence among different imaging modalities, the absence of the 
perfectly aligned training data and generic similarity function. Recently, 
several deep learning-based image registration methods have been proposed 
to address the difficulties in multimodal registration and achieved the 
state-of-the-art results over traditional methods in many tasks. As there 
is rapid adoption of deep learning-based medical image registration 
applications over the past few years, it is necessary to have a 
comprehensive summary and outlook. The main scope of this survey is to 
focus on the methodology and challenge of the deep learning-based 
multimodal registration methods. We first introduce the relevant deep 
learning-based multimodal registration proposed in the past few years and 
highlight the innovations and challenges for each method. Further, we 
propose possible future research directions and discuss how this field 
could be possibly moved forward to the next level.


Date:			Friday, 20 September 2019

Time:                  	10:00am - 12:00noon

Venue:                  Room 2303
                         Lifts 17/18

Committee Members:	Prof. Albert Chung (Supervisor)
 			Dr. Pedro Sander (Chairperson)
 			Dr. Qifeng Chen
 			Dr. Sai-Kit Yeung (ISD)


**** ALL are Welcome ****