Segmentation of magnetic resonance images for diagnostic applications.
Osteoarthritis (OA) and Bone Marrow Edema (BME) are very common diseases of the knee. They are characterised by the thinning of the joint cartilage and instances of abnormal bone respectively. Magnetic Resonance Imaging (MRI) has been increasing popular in characterization of the knee as it allows f...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/45431 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Osteoarthritis (OA) and Bone Marrow Edema (BME) are very common diseases of the knee. They are characterised by the thinning of the joint cartilage and instances of abnormal bone respectively. Magnetic Resonance Imaging (MRI) has been increasing popular in characterization of the knee as it allows for excellent soft tissue contrast and high-spatial resolution. Image segmentation is critical when analysing these MRI images in order to quantitatively determine the condition of the various components of the knee. In this project, we developed and evaluated a fully automatic image segmentation method for application to MRI Isotropic and Non-isotropic, PD-weighted and T2-FSw images of the knee joint. The structures of interest were the Femur, Femoral Cartilage, Tibia, Patella and Patella Cartilage. Two types of Deformable Registrations in Insight Toolkit (ITK) were used. This first was B-Spline Registration and the second was Free-Form Deformation (FFD) Registration. Mutual Information (MI) was used as similarity measure metric and Dice Similarity Coefficient (DSC) was used as an evaluation measure of the quality of the registration. |
|---|