Mathematical Sciences

Lund University

Title A Fully Automated Segmentation of Knee Bones and Cartilage Using Shape Context and Active Shape Models
Authors Behnaz Pirzamanbein
Full-text Available as PDF
Year 2012
Document type StudentPublicationsH2
Language eng
Abstract Swedish In this master&#39;s thesis a fully automated method is presented for seg-<br> menting bones and cartilage in magnetic resonance imaging (MRI) of the<br> knee. The knee joint is the most complex joint in the human body and<br> supports the weight of the whole body. This complexity and acute task of<br> the knee joint leads to a disabling disease called Osteoarthritis among the<br> adult population. The disease leads to loss of cartilage and torn cartilage<br> cannot be repaired unless surgical techniques are used. Therefore, one of<br> the important parts of nding the disease and planning the knee surgery<br> is to segment bones and cartilages in MRI.<br> The segmentation method is based on Statistical Shape Model (SSM)<br> and Active Shape Model (ASM) built from a MICCAI 2010 Grand chal-<br> lenge training database. First, all the data are represented by points and<br> faces. A Shape context algorithm is applied on 60 data sets to obtain<br> consistent landmarks. The mentioned consistent landmarks and Princi-<br> pal Component Analysis are used to build a Statistical Shape Model. The<br> resulting model is used to automatically segment femur and tibia bones<br> and femur and tibia cartilages with Active Shape model. The algorithm is<br> tested on the remaining 40 MRI data sets provided by the Grand challenge<br> 2010, and compared with six other submitted papers.