The morphological research on the construction of mandible finite element model using 2D cephalometric images

• Main Authors: Chih Hsun Lin, 林志勳
• Other Authors: Ching Tsorng Tsai
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/11950569158935701682

碩士 === 東海大學 === 資訊工程與科學系 === 92 === The traditional cephalometric analysis could only have a thorough observation of the front and back side symmetry of the mandible instead of the right and left side symmetry of the mandible. Such deficiency is incapable to point out the gradual influence on the balance development of mandible for the people who suffer from malocclusion as they try to find the suitable way to masticate. Finite Element method is widely used on the comparison morphology and medical diagnosis while the cephalometric analysis today is only on the 2D plane, rare on the 3D analysis Thus, we address the method to construct 3D finite element model by using 2 PA view and Lateral view cephalographs to make up the deficiency of the traditional cephalometric analysis. First, let the doctor point out the landmarks which are obviously shown on the PA view and Lateral view, then we can get the 3D coordinate values from these two 2D images landmarks. Later we use volume spline interpolation to compute the coordinate values of the other not easy to be seen landmarks, then we use OpenGL to construct 3D landmark model and further to get the 3D FEM model. At last we take the 2D and 3D cephalometric analysis and 2D and 3D finite element analysis into comparison to figure out the mandibular morphological changes. In the experiment, we temporarily disregard the image scaling problem and take the regular bite mandible as generic model as provided by China Medical University Pei Kang Hospital, then we lean the mandible slightly to left and adjust it to “Case i: Left bite”. According to the China Medial University, Department of Anatomy, the asymmetric mandible is referred as “Case ii”. We take X ray images from Cephalometry and compare the coordinates from volume spline interpolation and the real coordinates and we found out that the coordinates we got can reach as high as 90% accuracy, which shows that we can have a more accurate analysis on the morphological change of mandible.