散乱视频集的叙事式表达及编辑技术研究

项目名称： 散乱视频集的叙事式表达及编辑技术研究

项目编号： No.61472288

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 计算机科学学科

项目作者： 肖春霞

作者单位： 武汉大学

项目金额： 82万元

中文摘要： 本项目研究散乱视频集的叙事式表达及编辑，将某一地点或事件拍摄的杂乱无章、时空无序的视频集有效表达为一个完整的、时空连续的地点或事件，便于人快速理解视频集的主要内容。具体地，本项目将研究如下三个内容：（1）散乱视频集的时空一致定位及冗余视频过滤，为叙事式表达提供有代表性的素材。（2）散乱视频集的连接及相邻视频之间过渡，将定位、过滤后的视频集通过时空连接进行有效组织，并进行视觉愉悦的过渡。（3）散乱视频集多层次、多尺度叙事式表达，解决对叙事结果进行全方位、快速的浏览。本项目为上述研究内容提供了详细的研究方案，将散乱视频集的时空定位、过滤、连接、过渡、多层次表达等相关核心算法作为一个整体来思考，层层深入地进行研究。本项目拟在散乱视频集的叙事式表达的理论和方法、关键技术等方面展开系统、深入的研究，取得重要的创新性成果，提供高效的、符合人类视觉需求的视频叙事算法。

中文关键词： 散乱视频集；视频叙事；视频编辑；视频融合；纹理合成

英文摘要： This project is focusing on narrative processing of unconstructed video data. This project works to produce spatio-temporally continuously narrative results for the input unstructured large collections of video clips, which contain the same location, environment, or event, and make the video data easier to comprehend and explore. More specially, this project will work on the following three topics: （1） The global location estimation and filtering of the unstructured video set, which presents required representative videos for narrative. (2) Videos connection and transitions processing, this works to orginaize the unstructured video set, develop spatio-temporal transition between the neighboring videos, and make the narrative visually-pleasing for navigating. (3) Developing hierarchical and multi-scale video narrative, which makes it easy to explore the narrative comprehensively and in high efficiency. The project preset detailed research planning and techniques for these topics, and makes unified research scheme for the orientation, filtering, connection, transition, and multi-scale representation for the video set narrative. This project tries to carry out sophistical and systematic study on the theories,mythologies and critical techniques for the narrative processing of unconstructed video data, tries to achieve important and creative results, and works to present efficient, perception-pleasing video narrative algorithms.

英文关键词： Unstructured Video Set;Video narrative;Video editing;Video composition;Texture systhesis