ABSTRACT
Thanks to the rapid development of modern medical devices and the use of digital systems, more and more medical images are being generated. This has led to an increase in the demand for automatic methods to index, compare, analyze and annotate them. Until 2005, automatic categorization of medical images was often restricted to a small number of classes. The ImageCLEF medical image annotation challenge was born in this scenario, proposing a task reflecting real life constraints of content based image classification in medical applications. In this chapter we report about our experience first as participants, then as co-organizers. This research activity started in 2007, supported by a one-year IM2 fellowship. By leveraging over the initial IM2 support, in 2008 a 4-year project started (EMMA, Enhanced Multimodal Medical data Access), sponsored by the Hasler Foundation. Since 2009, the author has been an ImageCLEF task organizer, respectively for the medical annotation and robot vision tasks. Since 2013, she is the main organizer of ImageCLEF.
