Video Retrieval using Search and Browsing

We describe our experiments for the shot boundary detection, high-level feature extraction, search and story boundary detection tasks. In the shot boundary detection task, we employ our proven method based on the calculation of distances between colour histograms of frames over a range of timescales. In the feature detection task we confined our effort to the basketball feature. For the search task, we test a complete redesign of last year’s interface. Like last year, contentbased search is complemented with NN browsing functionality. We compare performance of two interactive runs, one including search and browsing, the other including search only. Overall search performance falls behind that of last year, a result which we would like to attribute to the stronger emphasis on motion in this year’s search task.