Selected Readings in Vision and Graphics
edited by Luc Van Gool, Gábor Székely, Markus Gross, Bernt Schiele

Volume 43

Christoph Spuhler

Interactive Centerline Finding in Complex Tubular Structures

First edition 2006, XII, 160 pages, € 64,00. ISBN 3-86628-114-5

Medical imaging has become increasingly important in recent years for screening, diagnosis and surgical planning. Improvements in scanning techniques in CT and MRI yield large 3D datasets of increasing resolution and quality. This has lead to new medical approaches such as virtual endoscopy.

With the advent of virtual endoscopy, path planning through complex organs, such as the small bowel, have become a necessity. Without efficient and accurate path planning virtual endoscopy becomes difficult and is often not feasible.

Until now only one efficient path planning procedure has existed for the small bowel, using haptic feedback for guidance. We have developed a new partial line extraction approach for complex tubular structures, which allows us to improve the path extraction. This is to our knowledge the first attempt for a semi-automatic line extraction in the small bowel. Using this technique, we are able to automatically extract a correct centerline in the small bowel for around 85% of the organ.

Using this line extraction as a basis, we have developed two approaches for path planning using either haptic feedback or the selection of segments using a keyboard. With the help of a pre-extracted partial centerline the haptic guidance can be improved. The second approach using segment selection offers an efficient alternative, which can be used in a clinical setting without the need for a haptic device or special training. The line extraction can also be used to efficiently extract a centerline in the colon.

We have developed a workstation which incorporates our path-planning approaches and allows a fast examination of the small bowel and the localization of polyps. We have performed trials in a clinical setting to verify the suitability of the workstation for polyp detection and localization. These show our approach to be fast and accurate.

Christoph Spuhler studied Electrical Engineering at the University of Arizona, receiving his B.Sc. in 2001. During this time he spent one year on exchange at the University of Stuttgart. In 2002 he obtained his M.Sc. in Computer Engineering from the University of Rochester in Rochester, New York, where his focus was on VLSI. In 2002 he joined the Computer Vision Laboratory at the ETH Zürich. Here he developed semi-automatic 3D line extraction methods in CT images for virtual endoscopy in the small bowel using haptics and automatic extraction algorithms. The work was performed under the scope of the European TOUCH-HapSys project. In 2006, he finished his doctoral thesis and was awarded the Ph.D. degree (Doctor of Sciences) from the ETH Zürich.

Keywords: virtual endoscopy, centerline extraction, small bowel segmentation, semi-automatic segmentation, haptics.

Schlüsselwörter: Virtuelle Endoskopie, Zentrallinien Berechnung, Dünndarm Segmentierung, Semi-automatische Segmentierung, Haptik.

Reihe " Selected Readings in Vision and Graphics " im Hartung-Gorre Verlag

Direkt bestellen bei / to order directly from: Hartung.Gorre@t-online.de