應用區域成長法於膝關節電腦斷層影像中骨組織分割之研究

Abstract

膝關節輔助檢查影像主要包含以下幾種：超音波、電腦斷層、核磁共振成像、X光、正子攝影。現今醫學影像技術日益進步，電腦斷層影像擁有多張切層影像資料，可以記錄膝關節各切面，並能進行重組顯示，以輔助醫生進行診斷及治療。不管在運動時發生膝關節受傷，或是因年齡老化而導致關節磨損，皆是需要進行膝關節檢查與治療，而本論文主要目的是膝關節骨組織影像分割，並將獲得的骨組織影像進行三維重建，讓醫生能從不同角度的骨組織三維影像進行診斷，使診斷效果更為精確。 本論文貢獻有以下三點：第一，本論文之系統可區分膝蓋部份與非膝蓋部份；第二，提出一個斷點修補方法，提升骨組織分割之效果。第三，基於貢獻1與貢獻2設計一個膝關節骨組織分割的系統雛形，並實驗證實本系統能達到良好的分割效果。

Ultrasound, computed tomography and magnetic resonance imaging are often use to examine knee joints. Owing to new technology, computed tomography provide more slices in shorter period of time. Moreover, it can not only reconstruct sequential image of bones in 3 dimensions, but record different planes to help diagnosis and therapy for doctors. Bone Segmentation is becoming increasingly important under different circumstances, such as, injuries during exercises, or the osteoarthritis caused by the degeneration. In this thesis, we design a system which can segment bone of knee joint image and reconstruct to get accuracy result. There are three contributions of this thesis. First, we propose a method can automatic distinguish the part of knee joints or the others. Second, we propose a improvement method to solve break point, it will be helpful to promote segmentation result. Third, based on above two contributions we complete the prototype of system in bone segmentation and reconstruction and prove our method can get good result.