The goal of this challenge is to compare fully automatic methods for segmenting anatomical structures in head and neck CT data used in radiotherapy planning. This year, bilateral parotid glands are offered for the competition. They represent important organs at risk, which need to be accurately delineated in preparation for radiation treatment.

The evaluation of the segmentation accuracy will be done by slice-by-slice comparison with manual delineations using dedicated validation methodology, one specific criterion being the assessment of the amount of the required manual corrections. Data: The data for this challenge is provided by Princess Margaret Hospital, Toronto. It consists of 25 CT datasets with the typical resolution of 1x1x2 mm and represents a wide variety of patient body types. Ground truth segmentations have been created by a clinical expert using standard manual delineation tools. Ten datasets are available for training with provided expert segmentation, to be used for fine-tuning the algorithms; eight datasets will be used for testing them against hidden expert segmentation and the last seven unseen datasets will be used onsite for the competition during the workshop. The format used for the provided data and to be used for submitted results is detailed in the Download/Submit section (binary .raw + ASCII .mhd header). Evaluation: The evaluation of the segmentation accuracy will be done by slice-by-slice comparison with manual delineations using dedicated validation methodology, one specific criterion being the assessment of the amount of the required manual corrections The content of this website is copyrighted © 2010 Vladimir Pekar, John Kim, Stéphane Allaire, Arish Qazi, David A. Jaffray