The development of an integrated computer based system for knee joint related clinical diagnosis, pre-surgical planning and post-operative monitoring

• Main Author: Ho, Jiahui
• Other Authors: Poh Chueh Loo
• Format: Theses and Dissertations
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Bioengineering
• Online Access: https://hdl.handle.net/10356/42377
• Institution: Nanyang Technological University

Knee joint is the largest joint within the human body and is most prone to injuries. This thesis presents the works that were carried out as part of a major research to develop an integrated system for knee joint related clinical diagnosis, pre-surgical planning and post-operative monitoring. The research is divided into three phases. Phase one involves in the creation of image segmentation modules for specific tissues at the joint, while phase two focuses on reconstruction of three-dimensional models from segmented images. Phase three dwells in the development of clinical based application modules. This project focuses on phase one of the development of the proposed integrated system. The specific aims of the project are to develop semi-automated segmentation techniques for anterior cruciate ligament (ACL) and menisci of the knee joint in magnetic resonance (MR) images, because these two tissues are most prone to damages when a knee joint is injured. Achieving these aims will set the foundation for future developments of this integrated system. The first presented work is the profile studies of knee joint related MR images. This study revealed a reproducible landmark in the vertical intensity profile which can aid computer based systems in locating the menisci automatically in MR images. In the second and third works, methodologies to the innovated semi-automated segmentation programs of ACL and meniscus are presented respectively. Both programs relied on morphological operations to locate the anatomical structures and active contour to perform the segmentation. Proton density weighted isotropic MR images were acquired and used to test the programs developed. Validation studies were conducted and the methodologies developed were proven to be feasible in achieving the segmentation tasks.