欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
现有的视频摘要技术缺乏一个统一、可扩充的视频摘要模型。针对这一缺陷,提出了实体-描述-效用模型（简称EDU模型）,该模型从视频实体出发,经过描述得到效用,并最终根据效用来生成视频摘要。该模型拥有一个作为中间信息共享平台的描述体系,以及基于描述体系的效用函数,具有可扩展性。对EDU模型的概念和结构进行了详细阐述,并根据该模型,提出了新闻视频故事摘要生成的方法。实验结果表明,该方法具有令人满意的效果。
英文摘要：
The current video summarization technique is short of a normal and expandable summarization model. To solve this problem, a model called EDU, which was shortened from the name of entity, description and utility model, was proposed. The model starts from video entity, gets utilities after descriptions, and generates video summarization by the utilities. As a general expandable model, EDU has a description scheme to save half-processed information, and a utility function based on descriptions. The concepts and structures of EDU model were described, and a method of news story summarization based on this model was produced. The experiment proves the effectiveness of the method.
同期刊论文项目
同项目期刊论文
期刊信息
位置：