Fully Automatic Carotid Artery Tracking and Segmentation

MPhil Thesis Defence


Title: "Fully Automatic Carotid Artery Tracking and Segmentation"

By

Mr. Evan HANN


Abstract

The carotid arteries comprises the common carotid artery (CCA), the 
internal carotid artery (ICA), and the external carotid artery (ECA). They 
are the major arteries for supplying blood to the brain, the face, and the 
neck. Unfortunately, these arteries are susceptible to stenosis, which is 
the narrowing of the vessel lumen due to accumulation of plaque. As the 
condition worsens, it can lead to life-threatening stroke - the world’s 
second leading cause of deaths according to World Health Organisation 
(WHO). Quantitative medical assessments such as 3D segmentation of the 
carotid arteries are necessary for monitoring this highly-concerning 
disease.

In the light of this, researchers have been keenly interested in 
automating 3D segmentation of the carotid arteries in place of tedious and 
laborious manual segmentation. A grand challenge of (semi) automatic 
segmentation of carotid artery lumen was held during the MICCAI conference 
in 2009. A total of 8 groups of participants submitted their works to the 
challenge. However, only 1 of the submissions was fully-automatic; the 
rest still require user-interventions such as seedings, albeit minimal. 
Developing a fully-automatic method for segmenting the carotid arteries is 
much more challenging, due to the presence of pathologies, the anatomical 
variabilities, and the imaging variabilities.

In this thesis, we have proposed a novel fullyautomatic method for tracking and 
segmenting the carotid arteries. The proposed method utilizes a wide range of 
image processing and analysis techniques including automated localization, bone 
elimination, vesselness filtering, vessel tracking, and vessel segmentation. 
None of the processes require humaninterventions. In experiment, the proposed 
method has demonstrated high success rate (48/56) in carotid artery tracking, 
outperforming the state-of-the-art fully-automatic method.


Date:			Thursday, 22 September 2016

Time:			9:00am - 11:00am

Venue:			Room 1504
 			Lifts 25/26

Committee Members:	Prof. Albert Chung (Supervisor)
 			Prof. Chiew-Lan Tai (Chairperson)
 			Dr. Pedro Sander


**** ALL are Welcome ****