Font Size: a A A

Multi-segmenatation Image Associated Registration Technology For Pedicel Screw Insertion Based On Structured Light Scanning

Posted on:2010-05-08Degree:MasterType:Thesis
Country:ChinaCandidate:X H LinFull Text:PDF
GTID:2178360278462158Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
Because of the complexity of anatomy of spine vertebrae, there is a significant rate of incorrect placement of transpedicular screws when using conventional surgical method. The registration method of preoperative CT image and intraoperative structured light scanning image is presented in this paper, and then computer-assisted pedicle screw insertion is achieved to improve the surgical accuracy.Firstly, a CT scan of the whole area of the spine to be operated on is obtained, and then point cloud data of bone surface are extraced using iso-surface method. Intraoperatively, the spine vetebrae is exposed and scanned by using structured light scanning, and both data are post processed using Mimics and CATIA softwares.To resigtrate CT image and structured scanning image, a registration method combined pre-registration and multi-suface ICP algorithm is presented. Firstly, the initial coordination transformation is obtained according to the curvature and normal vector of spine surface. Then a more pricise registration is computed using multi-surface ICP algorithm based on the initial transformation.To evaluate the accuracy and reliablity of the registration method, an experiment is conducted on a plastic phantom of spine. It shows that the registration accuracy of 0.5mm can be obtained even when a small portion of the posterior vertebral surface is used for registration. Also, the submillimetirc registration accuracy can be achieved even when the noises are imposed on the structured light image. Furthermore, the threshold value of CT image is optimized to increase the registration accuracy.The following animal experiments conducted on a cow spine also prove the feasibility of the registration method, and about 0.5mm registration accuracy is achieved. The range of threshold value is also optimized, which can be a reference for the in vivo surgery.Finally, this paper uses MFC-based OpenGL programming to faciliate the reliazation of registration algorithm, display of registration result and the interactive operation in surgery.
Keywords/Search Tags:Computer-assisted orthopaedic surgery, Registration, Structured light scanning, Image segmentation, ICP algorithm
PDF Full Text Request
Related items