Date of Award
Doctor of Philosophy
Industrial and Manufacturing Systems Engineering
Matthew C. Frank
Automated separation of reconstructed bone joints from 3D medical images is a challenging task due to surrounding soft tissue and adjacent bones that can affect the clarity of bone boundaries. Existing approaches typically require human intervention to correct improper results of segmentation before the joint model is reconstructed. This dissertation presents a new methodology for separating bone joint models using a completely automated approach. Rather than trying to offer a solution for segmenting medical images, the proposed method first allows errors in the reconstructed model and later removes these errors without the help of a medical expert or technician. This method utilizes known anatomical information from a generic CAD model, which is a properly generated model of the anatomy of a similar human subject, with regard to age, gender, height, etc. The intent is to aid in the separation of bones in the joint areas by comparing the reconstructed model that might contain errors to the generic model which has individual bones separated properly. The human hip joint is employed as an example of algorithm implementation in this dissertation. The proposed method is a general approach that should be adequately flexible to extend to other type of joints such as knee and elbow.
Boonsuk, Wutthigrai, "Automated separation of bone joint structures for medical image reconstruction" (2009). Graduate Theses and Dissertations. 10843.