Establishing the Utility of NonText Search for News Video Retrieval with Real World Users.pdfVIP

Establishing the Utility of Non-Text Search for News Video Retrieval with Real World Users Michael G. Christel School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 USA 1-412-268-7799 christel@ ABSTRACT Automated tool support in combination with human manipulation TRECVID participants have enjoyed consistent success using and interpretation offer tremendous leverage in addressing the storyboard interfaces for shot-based retrieval, as measured by challenge of video information retrieval. Without automated TRECVID interactive search mean average precision (MAP). tools, the human user is swamped with too many possibilities as However, much is lost by only looking at MAP, and especially by the quantity and diversity of video accessible on the Web neglecting to bring in representatives of the target user proliferate. Ignoring the human user, though, is a mistake. Fully communities to conduct such tasks. This paper reports on the use automated systems involving no human user have consistently and of within-subjects experiments to reduce subject variability and significantly underperformed compared to interactive human-in- emphasize the examination o