Exploring the utility of fast-forward surrogates for bbc rushes

Abstract

This paper discusses in detail our approaches for producing the video summaries submitted to the TRECVID 2008 BBC rushes summarization task, including the baseline method. Empirical work produced during and after the TRECVID 2007 rushes summarization task gave strong evidence that a simple 50x method (sampling every 50 th frame) provides excellent coverage (text inclusion performance). Our submissions for TRECVID 2008 investigated the effects of junk frame removal, including a comprehensible audio track, and emphasizing pans and zooms when backfilling to reclaim the space removed with the noise shots from the original 50x set. Results show that 50x based methods provide excellent coverage as expected. There were limited effects for the other strategies to improve user satisfaction, with the discussion providing some insights for future video summary development and evaluation work.

Automated Metadata in Multimedia Information Systems: Creation, Refinement, Use in Surrogates, and Evaluation

“…For example, the fast-forward surrogate of playing a BBC rushes source video at 50× to produce a 2% summary leverages no context aside from the removal of "junk" frames (color bars, scene clappers) [92]. Multimedia surrogates representing a single multimedia document or sets of documents also can be improved through interactive user control and user context.…”

“…The assessment found that the skim was more effective if composed of longer 5-second phrasealigned audio chunks than shorter 2.5-second chunks [1]. For BBC rush materials with much redundancy due to repeated takes, 50× worked nearly as well as 25× and offered much greater efficiency [91], with the contribution of an audio component to the playable 50× summary still an open area of investigation [92]. Playable video summaries have received renewed attention from the TRECVID community in its BBC Rushes Summarization Task, looking at indicative and informative summaries that have at most 4% of the source length in 2007, and 2% of the source length in 2008.…”

A Survey on Visual Content-Based Video Indexing and Retrieval

“…They match and mosaick the extracted key frames to reduce content redundancy and to build compact visual summaries. Christel et al [276] have constructed a baseline rushes summarization system at TRECVid 2008 [272]. This baseline method simply presents the entire video at 50×normal speed.…”

The trecvid 2008 BBC rushes summarization evaluation

“…The baseline simply presented the entire video at 50× normal speed -a strategy arrived at after study of various alternatives [6]. The baseline was mute with no audio whatsoever.…”