Statistical shape model of the knee joint

• Main Author: Lim, Cheng Wei.
• Other Authors: Poh Chueh Loo
• Format: Final Year Project
• Language: English
• Published: 2010
• Subjects: DRNTU::Engineering::Bioengineering
• Online Access: http://hdl.handle.net/10356/39610
• Institution: Nanyang Technological University

The knee joint is a complex joint which has been a popular subject for research. Every year, more people succumb to knee injuries, hence the need to conduct a study of the knee. In this project, the aim is to devise a methodology to build an atlas of the knee joint. The images are first segmented to extract the structure of interest. Principal Axes Transformation (PAT) is then performed to rigidly register the images. After PAT, Statistical Shape Model (SSM) is used to generate a mean shape which constitutes the atlas. From SSM, we are also able to generate observe variations in shapes by changing shape parameters. The study is conducted on MRI images of the femur bone. Artificially created images allowed the understanding of how to generate an SSM. SSM is then conducted on 10 subjects: 5 male and 5 female. A comparison between male and female subjects revealed that male medial & lateral condyles, and female medial condyle varied at approximately the same rate. However, variation in the female lateral condyle is significantly higher than the other condyles. Atlases of ACL attachment sites of the femur and the tibia are also generated using SSM to demonstrate that SSM can be used for other knee structures. This project serves as a preliminary study for building an atlas of the knee joint. Future work may involve building a 3D atlas of the knee and automating the atlas building process.