Welcome to the SKI10 website. The goal of SKI10 was to compare different algorithms for cartilage and bone segmentation from knee MRI data. Knee cartilage segmentation is a clinically relevant segmentation problem that has gained considerable importance in recent years. Among others, it is used to quantify cartilage deterioration for the diagnosis of Osteoarthritis and to optimize surgical planning of knee implants. See the SKI10 paper in the SKI10 Zenodo repository for further details.
SKI10 started out as one of the three competitions of the Grand Challenge Workshop 2010, organized in conjunction with the MICCAI 2010 conference.
Current status
SKI10 has been running from 2010 to 2018. During this time, 170 teams have registered for the challenge and tried to solve it. 15 teams have submitted results for 24 algorithms that are listed in the results table. In 2019, we were informed that we are no longer allowed to share the SKI10 dataset, which was kindly provided to us by Biomet, Inc., and the challenge is officially closed. Thank you to all participants, it was a great journey!
For further information, see the answers to frequently asked questions.
Who was organizing the challenge?
SKI10 was organized by the following people:
- Tobias Heimann, German Cancer Research Center, Heidelberg, Germany
- Bryan Morrison, Biomet Inc., Warsaw, IN
- Simon Warfield, Harvard Medical School, Boston, MA
- Martin Styner, University of North Carolina at Chapel Hill, NC
- Marc Niethammer, University of North Carolina at Chapel Hill, NC