Automatic Segmentation and Motion Tracking of Papillary and Trabecular Structures from Cardiac MRI

Project Code: 10336060

Faculty: Faculty of Medical and Health Sciences

Department: Anatomy

Main Supervisor: Avan Suinesiaputra

Application open date: 28 Aug 2015

Application deadline:

Enrolment information: NZ Citizens, NZ Permanent Residents, International

Introduction

In some patient groups, such as hypertropic cardiomyopathy and congenital heart defects, the assesment of ventricular mass and volumes is significantly affected by the presence of papillary muscles and trabeculae. However, in the current clinical analysis of cardiac MRI, papillary muscles and trabeculae are usually excluded from the ventricle. The reason is the difficulty to delineating these irregular structures, which creates a low reproducibility of the analysis. This has been a challenge by the medical image analysis community to develop a fast, accurate and highly reproducible automated segmentation algorithm.

What we are looking for in a successful applicant

We want a student with a strong interest in developing a computer algorithm for image segmentation and motion tracking. Ideally, this would be a computer science or bioengineering student.

Objective
  • Develop an algorithm to segment papillary muscles and trabeculae from cardiac MRI using either Matlab or C/C++.
  • Extend the algorithm to track motion of papillary muscles and trabeculae throughout the cardiac cycle with the constraint of a constant mass.
  • Test and validate the algorithm in 10 cases of patients with congenital heart defects where we have manual contours from 2 observers.
  • Prepare a journal publication.
Other information
Related opportunities