粉碎性骨折虛擬復位模擬與量測研究：利用三維定位與三維列印混合式技術;Virtual Bone Reduction and Measurement of Comminuted Bone Fracture: Hybrid use of 3D Registration and 3D Printing Technologies

NCU Institutional Repository > 工學院 > 機械工程研究所 > 博碩士論文 > Item 987654321/81574

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/81574

题名:	粉碎性骨折虛擬復位模擬與量測研究：利用三維定位與三維列印混合式技術;Virtual Bone Reduction and Measurement of Comminuted Bone Fracture: Hybrid use of 3D Registration and 3D Printing Technologies
作者:	Irwansyah;Irwansyah
贡献者:	機械工程學系
关键词:	虛擬骨骼復位;3D 配對;組裝 3D 列印物件;骨骼形態學;Virtual bone reduction;3D registration;Assembly 3D-printed parts;Bone morphology
日期:	2019-05-20
上传时间:	2019-09-03 16:27:04 (UTC+8)
出版者:	國立中央大學
摘要:	藉由把傳統手繪草圖電腦化的過程中，術前規劃系統已經到達新的里程碑，3D 術前規劃系統幫助外科醫生建立清晰的 3D 骨頭模型並且增加醫生對真實破損骨頭的認知，為了修復破損的骨頭回復到原本的骨頭結構位置，人機介面分析工具(Computer mouse, keyboard and haptic devices)常被用來引導破裂的骨頭碎片，此外，利用多對點約束來配對骨頭碎片為大眾的對應方式，然而，這些研究通常是複雜且不直觀，而且常重複運算造成時間損失，特別是配對兩個解析度不高的骨頭碎片點資料中，很難精準的回覆骨頭碎片。本研究基於模板化、解剖學指標與破裂線選擇上的考量，來解決配對對應點的限制，半自動點配對方法被用來轉換骨頭碎片到相對應的目標，其演算法是根據單數值分解，為了量化建議的配對點結果，需計算位置與旋轉的均方根誤差，目前的配對方法已經在不同層面的錯誤中，完成骨頭復位，其結果與其他研究並沒有太大的差別，在不同的破碎案例中，每個校正方法各有優缺點，因此，以上討論的配對物件將會幫助使用者在辨識配對點和虛擬破碎骨頭復位的過程中減少手術時間。組裝和驗證 3D 列印之破碎骨頭的研究為確保在電腦建模中的 3D 骨頭幾何模型與實際上 3D 列印出來的模型一樣，網格的完整性在轉換 3D 模型至可列印之檔案是很重要的，辨識和移除重疊網的演算法，可以用來增進 3D 骨頭模型，在每一個模型中，擁有最大均方根誤差亦為百分誤差之面編號是需要被計算的，結果顯示 3D 列印出來之骨頭可以簡單的組裝而且在破裂面上只有細微的裂縫，而這些裂縫是在可接受的誤差中。除了評估 3D 列印出來之骨頭，也針對骨頭的型態做探討，目前針對正常骨頭型態的參數已經可以被計算出來，其結果被儲存在資料庫中供後續作為虛擬骨頭復位過程的參考資料，在未來將會針對骨頭復位方面做演算法的改進，並且持續增進辨識與移除網格瑕疵功能，以增進骨頭型態的自動化計算，衍生之計畫為藉由通過 3D 列印與自動骨裂復位技術，增進術前規劃系統，這項研究可避免在過程中使用錯誤嘗試法並準確地復位骨頭，減少手術時間與減少輻射吸收。;Three-dimensional (3D) preoperative planning system has gained important and new milestone by transforming the traditional manual sketching to computerization. The system assists surgeons to generate the bone model with clear visibility and increase the understanding of actual damaged bones. In order to recover the bone fractures back into their anatomic original pose, user interactive approach is commonly used as a navigation tool. Multi-paired point is the popular registration constraints to match the bone fragments. However, this method is tedious, and repetitive tasks, specifically when it is necessary to accurately relocate bone fragments due to difficulties in determining matching points in two unclear visibility of given fragments. Templated based, landmark point based and fracture lines based registration are studied and evaluated to solve the limitation of paired points based registration. A semi-automatic bone reduction was proposed to match the fragment with it counterparts. To quantify the results of the proposed registration, displacement errors in fragment position were calculated. The proposed registrations have accomplished bone reduction with various level errors. The results were not significantly different from other approaches. Each of the registration has its own strength and weakness that might be useful for different types of fracture cases. Hence, introducing the discussed types of registration features will assist the user in recognizing paired points and in reducing operation time during the virtual fractured bone reduction process. The assembly and verification of 3D-printed bone fractures were studied to ensure the geometric result of 3D bone modelling in the computer process similar to 3D-printed physical parts. To transform the 3D model to be a file ready for printing, mesh correction and removal of common errors are necessary. The algorithm for detecting and removing of overlapping meshes were performed to improve 3D bone models. Facets number, maximum and RMS errors as well as percentage difference of the overlapping meshes on each bone model are calculated. The results show that 3D-printed bone parts could be easily assembled and only small gap errors were noted on the fracture surface regions within an acceptable range of deviation. Beyond the physical geometry of 3D-printed assessment, bone morphological parameters also evaluated. In present, the normal bone morphology parameters were computationally measured. The results are stored in database and used as the reference to evaluate the virtual bone reduction. In the future, we will improve our algorithm for supporting bone reduction, detecting and removing mesh defects, and enhancing automatic bone morphology measurement. Extending study was also considered by elaborating virtual bone reduction with 3D-printing and robotic technology to encourage the preoperative planning system. It may be beneficial to avoid the trial-and-error process and accurate bone reduction.
显示于类别:	[機械工程研究所] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	160	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....