CT Image analysis of proximal ulna for improved implant design.
Tan, Kenneth Yong Long.
Date of Issue2013
School of Chemical and Biomedical Engineering
Understanding the morphology of the proximal ulna using Computed Tomography (CT) scan is important in estimating the shape and size of elbow for improved implant design for the patients. The current methods are unable to accurately detect the canal, especially at locations close to the proximal ulna. The objective of this project is to determine the intra-medullary dimensions of the ulna by developing more advanced image processing technique, texture-based segmentation. MaZda, a computer software for calculation of texture parameters/features is used to perform CT image analysis of the ulna. CT scans of 19 proximal ulna are analyzed using a cross validation process. Analysis is done at 2R (location where simple methods are known to fail) to test the suggested methodology. Dice’s Coefficient (DSC) will be used to show the superiority of the developed method at this location. Our findings have proven that the suggested methodology gives a better result of an average DSC of 0.751, while the simple thresholding based on Hounsfield units (current method) gives an average DSC of 0.48.
Final Year Project (FYP)
Nanyang Technological University