TRECVID 2010 Experiments at Dublin City University

This year the DCU-CLARITY-iAD team participated in the both the instance search and interactive known-item search tasks of TRECVid 2010. For our instance search submission, we used classifiers to search for candidate objects in each keyframe. This was achieved by a coarse-to-fine search based on a hierarchical representation of the regions in the keyframes. Our results proved inconclusive, but we believe the method warrants further investigation. The 2010 interactive search task at TRECVid represents a number of firsts for the community and for our team it represents the first time that many of our team has participated in TRECVid. Our approach this year was to develop a simple and intuitive system which we felt could be used by video information seeking specialists and complete novices alike. To this end we have developed our 2010 TRECVid KIS system on an Apple iPad, the iPad is a new tablet computer developed by Apple, it represents a lean-back, relaxed and easy to use computer, likewise our search engine was designed to be easy to use by all users. Our underlying search engine allows for the three commonly used video search methods: text search, concept search and image search. For our experiments we compared the performance of our system when used by standard users from our research group versus novices with no technical expertise. Our results show that the two groups gave identical performance in terms of mean elapsed time.