FEATURE BASED ROBUST NON-RIGID IMAGE REGISTRATION IN SPATIAL AND FREQUENCY DOMAINS

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "FEATURE BASED ROBUST NON-RIGID IMAGE
REGISTRATION IN SPATIAL AND FREQUENCY DOMAINS"

By

Mr. Shu Liao


Abstract

Non-rigid image registration plays an important role in medical image 
analysis, disease diagnosis and statistical parametric mapping. In this 
thesis, we particularly focus on developing novel features for robust 
image registration and designing an efficient evaluation protocol to 
measure the robustness and discriminant power of the features.

First, in the spatial domain, a new image feature called the uniform 
spherical region descriptor (USRD) is proposed. The USRD feature is 
rotation and monotonic gray-level transformation invariant, and is also 
computationally efficient. Each voxel is represented by its own USRD 
feature signature. The USRD feature is integrated with the Markov random 
field labeling framework for image registration. Second, we propose the 
symmetric alpha stable (SαS) filters to extract image features in the 
frequency domain. The SαS filters are proposed because the energy 
spectrums of brain MR images often exhibit non-Gaussian heavy-tail 
behaviors which cannot be satisfactorily modeled by the conventional Gabor 
filters. The conventional Gabor filter is a special case of the SαS 
filters. The maximum response orientation criterion is designed to xiii 
make the SαS feature rotation invariant. The SαS feature is integrated 
with the subvolume deformation model in the registration process. 
Moreover, in this thesis, we propose the Fisher's separation criterion 
(FSC) protocol which can directly evaluate the discriminant power of 
various types of features.

Finally, a multi-layer framework is proposed to extract features from 
input images from different views. The proposed methods are evaluated by 
performing non-rigid registration experiments. The proposed methods are 
also compared with several state-of-the-art registration approaches. It is 
demonstrated that the proposed methods consistently achieve the highest 
registration accuracies among all the compared methods, which is matched 
with the results obtained from the proposed FSC evaluation protocol.


Date:			Thursday, 22 July 2010

Time:			2:00pm – 4:00pm

Venue:			Room 3501
 			Lifts 25/26

Chairman:		Prof. Alexis Lau (CIVL)

Committee Members:	Prof. Albert Chung (Supervisor)
 			Prof. Pedro Sander
 			Prof. Chi-Keung Tang
                     	Prof. Oscar Au (ECE)
                        	Prof. Pheng-Ann Heng (Comp. Sci. & Engg., CUHK)


**** ALL are Welcome ****