腫瘤偵測與顱顏骨骼重建; Tumor Detection and Craniofacial Implant Reconstruction

NCUIR > Engineering College > Graduate Institute of Mechanical Engineering > Electronic Thesis & Dissertation > Item 987654321/1842

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/1842

Title:	腫瘤偵測與顱顏骨骼重建;Tumor Detection and Craniofacial Implant Reconstruction
Authors:	徐榮祥;Jung-Shian Hsu
Contributors:	機械工程研究所
Keywords:	腫瘤;邊界偵測;顱顏重建;超音波影像;Tumor;Craniofacial;Boundary detection;rapid protyping machine
Date:	2002-01-09
Issue Date:	2009-09-21 11:32:52 (UTC+8)
Publisher:	國立中央大學圖書館
Abstract:	本論文題出一個新的顱顏重建的方法解決重建的問題降低手術所須要的時間,以類神經網路預測病灶區的外形並於臨床應用獲得良好的結果另外本論文亦提出方有效的超音波偵測乳房腫瘤外型並用於腫瘤良惡性的判斷 Traditionally, plastic surgeons reconstruct craniofacial defects according to their clinic experience while operation is in progress. It is time-consuming to make the implant and the hand-made implant is usually difficult to well match the defect. The purpose of this research is to propose a method to improve the effect and efficiency of traditional operation. In this study, the orthogonal neural network is applied to predict the surface model of the defect and then the Marching Cube Algorithm is applied to reconstruct the 3D defect implant model. A rapid prototyping machine can accurately produce the geometric of patient-specific implants in acrylic resin. Two clinic cases with either a forehead defect or a large skull defect are given to evaluate the performance of the proposed method. The results show that the reconstructed implants fit into the defects well . A method for tumor boundary detection and a procedure for the diagnosis of breast tumor are also presented. The grey level projection distribution of the ROI is adopted to determine the seed point and threshold value of the tumor. Then the tumor boundary can be determined by searching from the seed point and by using the region growth method. After the tumor boundary of each image slice has been determined, the tumor size and spatial position can be calculated accurately. The shape and margin of the detected tumor boundary can also be used to assist the prediction of breast tumor attributes. The method has been applied to detect the breast tumor boundary from sonograms and brain tumor boundary from CT image slices. The results of clinic tests show that the computer generated tumor boundary matches well with the subjective judgement of an experienced breast tumor expert and a neurosurgeon. In this study, fifty-four breast sonograms are analysed. In comparison with physician judgement, twenty-three cases reach 100% similarity. Fifteen cases reach 90% similarity and eleven cases reach 80%. However, one case only reaches 70% and four cases are different from the physician judgement.
Appears in Collections:	[Graduate Institute of Mechanical Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Size	Format

社群 sharing

Loading...