DOI: 10.1117/12.373569
关键词: Set (psychology) 、 Video tracking 、 Information extraction 、 Video processing 、 Computer science 、 Video editing 、 Automatic summarization 、 Multimedia 、 Video production 、 Variation (game tree)
摘要: An increasing number of people own and use camcorders to make videos that capture their experiences document lives. These easily add up many hours material. Oddly, most them are put into a storage box never touched or watched again. The reasons for this manifold. Firstly, the raw video material is unedited, therefore long-winded lacking visually appealing effects. Video editing would help, but, it still too time-consuming; rarely find time do it. Secondly, watching same tape more than few times can be boring, since lacks any variation surprise during playback. Automatic abstracting algorithms provide method processing so users will want play often. However, existing automatic have been designed feature films, newscasts documentaries, thus inappropriate home footage in general. In paper, we present new generating amusing, variable abstracts automatically. They new, empirically motivated approach, also presented cluster time-stamped shots hierarchically meaningful units. Last but not least, propose simple natural extension way acquire - so-called on-the-fly annotations which allow completely set applications on as well enable better selective abstracts. Moreover, our restricted applied