Story units are extracted by general tempo analysis including tempos of audio and visual information in this research. Although many schemes have been proposed to successfully segment video data into shots using basic low-level features, how to group shots into meaningful units called story units is still a challenging problem. By focusing on a certain type of video such as sport or news, we can explore models with the specific application domain knowledge. For movie contents, many heuristic rules based on audiovisual clues have been proposed with limited success. We propose a method to extract story units using general tempo analysis. Experimental results are given to demonstrate the feasibility and efficiency of the proposed technique.
|頁（從 - 到）||46-54|
|期刊||Proceedings of SPIE - The International Society for Optical Engineering|
|出版狀態||已發佈 - 2003|
|事件||Internet Multimedia Management Systems IV - Orlando, FL, 美国|
持續時間: 2003 九月 9 → 2003 九月 11
ASJC Scopus subject areas