Home About UsPeopleEducationLinks



Location: CGVG > CVG > Research > Projects


Research - Projects

overview |  projects |  software & demos |  publications |  seminars  ]



Tracking of Spinal Motion in Video-fluoroscopy

Principle Investigator: Dr. K.-K.Y.Wong

Low back pain is one of the most common health disorders and its cost is enormous. There is a general consensus that the diagnosis and the treatment of low back pain can be aided by analysing spinal movement. Thus, spinal measurement techniques have been studied widely. At present, videofluoroscopic imaging provides an effective method of obtaining images for spinal motion analysis. Generally, landmarks of a moving vertebra will be extracted from videofluoroscopic video and will then be analysed. Landmarks are usually the corners of the moving vertebra and are usually extracted manually. Unfortunately, the analysis is difficult and time consuming due to the low quality of the videofluoroscopic images. Thus, automatic extraction of landmarks is studied in this project.

In this project, a method in which an active contour (or snake) will attach to the vertebrae automatically throughout the video sequence is proposed. Users only need to define landmark positions on the first videofluoroscopic image. The active contour formed from such landmarks will attach to the vertebra automatically in the remaining video images. This greatly reduces the effort of physicians in setting accurate landmarks of vertebra manually in every video frame. The reduction in human intervention means the reduction in error rate due to fatigue of the operator. Analysis on spinal motion can be done much more effectively and accurately.




Computer Vision Group, Department of Computer Science, The University of Hong Kong
Copyright © 2004 All rights reserved. Contact Us
Last Updated: May 15 2005