脊椎表面取點區域及路徑對於疊合精準度影響之分析

Abstract

本文旨在探討脊椎手術導航系統在建立醫療影像座標系及物理空間座標系時，運用較不同的概念去做疊合的步驟，找出在相同演算法下，能讓目前導航精度在進一步提升的辦法。以之前文獻中的經驗，加入形狀特徵的想法，配合每個不同特徵的專屬方式與區域的組合，經過模擬及實驗找出最好的實驗結果。實驗中所採用的疊合演算法為迭代最近點，而CT檔案以Marching cube的概念重建為三角網格，再配合本研究所提出之取點辦法，進行實驗。經過多次的實驗及比較證明利用形狀特徵劃分區域配合專屬取點方式可以提高疊合精度，且軸向特徵越明顯疊合效果越好。

This article aims to explore spinal surgical navigation system in the establishment of medical imaging coordinate system and physical space coordinate system, the use of different concepts run the registration step to find out at the method which can improve the current navigation accuracy with the same algorithms. In previous literature experience,then add shape feature ideas, each with unique characteristics and combinations of different areas, through simulation and experiment to find the best results. The experiment used an iterative closest point algorithm for registration, and the CT reconstruction to triangular mesh used the concept Marching cube, and then take the points with the approach proposed in this study, conducted experiments. After several experiments and demonstrate the use of relatively complex shape is distinguished by exclusive area take some ways to improve the accuracy of registration, and axially overlapping features more obvious the better.