- 无标题文档
查看论文信息

中文题名:

 教学视频摘要的自动生成技术研究    

姓名:

 江梦婷    

保密级别:

 公开    

学科代码:

 080714T    

学科专业:

 电子信息科学与技术    

学生类型:

 学士    

学位:

 理学学士    

学位年度:

 2013    

学校:

 北京师范大学    

校区:

 北京校区培养    

学院:

 信息科学与技术学院    

第一导师姓名:

 余乐军    

第一导师单位:

 北京师范大学信息科学与技术学院    

提交日期:

 2013-05-21    

答辩日期:

 2013-05-17    

外文题名:

 The teaching video automatic generation technology research    

中文关键词:

 教学视频摘要 ; 关键帧提取 ; 逐帧拼接    

中文摘要:
近年来,随着计算机多媒体技术的不断发展,教学视频也渐渐在现代教育领域中有了越来越广泛的应用。但是如今海量的教学视频导致用户在使用时存在很大的困惑,教学视频数据的内容丰富性和视频时长等问题使得用户对视频浏览的效率、方便性均存在较大不足。如何快速高效的将教学视频中的主要信息抽取出来,形成一个简洁明了的摘要文件是教学视频应用的关键技术点。教学视频摘要不仅可以使视频数据进行快速的传输,也提高了用户的学习效率。本文在已有的算法研究基础上,针对教学视频这块领域进行了一定的研究。其主要研究内容如下: 提出了基于采样和运动分析的关键帧提取技术。首先,对教学视频在一定的采样间隔上进行采样得到候选关键帧。然后,在采样的基础上对候选关键帧进行运动分析提取出形成教学视频摘要所需的关键帧。该技术结合了基于采样和基于运动矢量估计各自的优点,在一定程度上提高了关键帧提取的准确性,同时减少了关键帧的数据冗余度。 在关键帧提取的基础上,提出了基于逐帧拼接的教学视频摘要自动生成技术。在逐帧拼接的同时,判断待拼接关键帧是否属于同一个黑板内容。如果是拼接继续,反之则生成一个摘要图像,进行下一个摘要图像的拼接。
外文摘要:
In recent years, with the continuous development of computer multimedia technology, teaching videos gradually has been more widely used in the modern field of education. But now a flood of teaching videos lead to users a lot of confusion when people use them. The teaching videos have the rich data and continue a long time. There is a big shortage about the efficiency of video viewing and the convenience. How fast and efficient to form a concise summary file from the teaching videos has been the key technical points of the teaching video applications. Teaching video summary not only allows fast transmission of video data, but also improve the efficiency of the user's learning. In this paper, the research based on the existing algorithms and research in the areas of teaching video. The main contents are as follows: Put forward a key frame extraction technology which based on the sampling and analysis of motion. First, extract candidate key frame at a certain sampling interval from the teaching video. Then, do motion analysis frame-by-frame for the candidate key frame. Finally extract the key frame form teaching video which the summary required. This technique combines their respective advantages based on sampling and based on motion vector estimation. To some extent, improve the accuracy of the key frame extraction and reduce the data redundancy . Put forward a technology based on frame-by-frame splicing to form a summary of the teaching video. The technology based on the key frame extraction. Analyzing whether the key frame which to be splicing belongs to the same content on the blackboard or not when we do the frame-by-frame splicing. Finally make a summary of the content about the work and point out some follow-up work.
馆藏号:

 本071201/1320    

开放日期:

 2013-07-31    

无标题文档

   建议浏览器: 谷歌 360请用极速模式,双核浏览器请用极速模式