Quaero at TRECVID 2010: Semantic indexing

The Quaero group is a consortium of French organization working on Multimedia Indexing and Retrieval1. UJF-LIG and KIT participated to the semantic indexing task and UJF-LIG participated to the organization of this task. This paper describes these participations. For the semantic indexing task, a classical approach based on feature extraction, classification and hierarchical late fusion was used. Four runs were submitted corresponding to the use or not of genetic algorithmbased fusion and of two distinct fusion optimization methods. Both led to a small performance improvement and our best run has an infAP of 0.0485 (33/101). UJF-LIG also co-organized the task with the support of the Quaero programme while taking care of avoiding conflict between participation and organization. We defined a new version of the previous HLF detection task, now called semantic indexing. Two versions of the task were proposed with different numbers of concept to detect: 10 for the "light" version and 130 for the "full" version. We organized as in 2007-2009 the collaborative annotation for this task using an active learning approach. An improvement was made this year by the use of relations between concepts during the annotation process. We also assessed 30 additional concepts for the evaluations. 10 of them were included in the official TRECVID 2010 evaluation and 20 more were delivered later.