liu.seSearch for publications in DiVA
Change search
ReferencesLink to record
Permanent link

Direct link
Segmentation of Medical Image Volumes
Linköping University, Department of Electrical Engineering, Computer Vision. Linköping University, The Institute of Technology.
1997 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Segmentation is a process that separates objects in an image. In medical images, particularly image volumes, the field of application is wide. For example 3D visualisations of the anatomy could benefit enormously from segmentation. The aim of this thesis is to construct a segmentation tool.

The project consist three main parts. First, a survey of the actual need of segmentation in medical image volumes was carried out. Then a unique three-step model for a segmentation tool was implemented, tested and evaluated.

The first step of the segmentation tool is a seed-growing method that uses the intensity and an orientation tensor estimate to decide which voxels that are part of the project. The second step uses an active contour, a deformable “balloon”. The contour is shrunk to fit the segmented border from the first step, yielding a surface suitable for visualisation. The last step consists of letting the contour reshape according to the orientation tensor estimate.

The use evaluation establishes the usefulness of the tool. The model is flexible and well adapted to the users’ requests. For unclear objects the segmentation may fail, but the cause is mostly poor image quality. Even though much work remains to be done on the second and third part of the tool, the results are most promising.

Place, publisher, year, edition, pages
1997. , 71 p.
LiTH-ISY-Ex, 1864
National Category
Engineering and Technology
URN: urn:nbn:se:liu:diva-54357ISRN: n/aOAI: diva2:303019
Available from: 2010-03-10 Created: 2010-03-10 Last updated: 2010-03-30Bibliographically approved

Open Access in DiVA

fulltext(1051 kB)221 downloads
File information
File name FULLTEXT01.pdfFile size 1051 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Lundström, Claes
By organisation
Computer VisionThe Institute of Technology
Engineering and Technology

Search outside of DiVA

GoogleGoogle Scholar
Total: 221 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 212 hits
ReferencesLink to record
Permanent link

Direct link