Development of an automatic summary editing system for Audio-Visual materials

音像资料自动摘要编辑系统的开发

基本信息

批准号：
62580252
负责人：
KAWAFUCHI Satomi
金额：
$ 1.22万
依托单位：
Tokai University
依托单位国家：
日本
项目类别：
Grant-in-Aid for General Scientific Research (C)
财政年份：
1987
资助国家：
日本
起止时间：
1987 至 1989
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-62580252/
关键词：
Summary data database audio-visual resources University of the Air TV program Retrieval system image and sound 遠隔教育映像音影資料デ-タベ-ス映像教材サマリー PCM データベース静止画映像教材サマリー情報映像教材データベースキー画面

项目摘要

The authors have developed a data base for secondary information on visual aids and have made it available for public use through the use of university networks. This date base covers secondary information compiled from brochures which explain teaching materials. From the use's point of view, however, if graphics as well as the secondary information are presented, the necessary data could be retrieved with more accuracy. In particular, in the case of the university situation covered by this study, it is unrealistic for learners to look at and listen to all materials on available subjects in order to choose the appropriate subjects and then find the particular section of interest. For this reason, we are required to prepare summaries of the contents of visual aid materials.Such summaries resemble previews, many of which are animated and with sound. If they are applied to the data base, it would be realistic to combine still pictures with sound because the use of animation require a huge … More memory. This study analyzed the elements of educational programs at a university to determine guidelines for preparing summaries.Although the summaries should be prepared by staff who have experience in producing visual aids and teachers, this entails some difficulties for various reasons. It is also almost impossible to instruct others to prepare a large number of summaries (e.g., thousands) because of the tremendous workload this entails.Therefore, this study tried to evaluate and experiment with 1) the equal-interval method with a timer and 2) the cut-responsive method focusing on the changing of scenes and thereby preparing automatically summaries by these methods.The cut-responsive method is used to detect changes in the brightness of video signals and to detect and store them electronically. The result of the evaluation showed that when the sound was taken out 2 seconds before the change of cut (based on an analysis of the programs) and the scenes were recorded 3 seconds after the change of cuts, key scenes and key words could be extracted. Thus, it was found to be useful. Less

作者开发了一个关于视觉教具的二手信息数据库，并通过大学网络向公众开放，该数据库涵盖了从使用角度解释教材的小册子中编制的二手信息。，如果提供图形以及辅助信息，则可以更准确地检索必要的数据。特别是，就本研究涵盖的大学情况而言，学习者查看和收听所有材料是不现实的。可用的科目以便选择出于这个原因，我们需要准备视觉辅助材料的内容摘要。这样的摘要类似于预览，其中许多是动画的并且如果应用于的话，则带有声音。数据库中，将静态图片与声音结合起来是现实的，因为动画的使用需要巨大的内存。这项研究分析了大学教育课程的要素，以确定准备摘要的指导方针。尽管摘要应该由以下人员准备。具有视觉制作经验的工作人员由于各种原因，这会带来一些困难，而且由于工作量巨大，指导其他人准备大量（例如数千）摘要也几乎是不可能的。因此，本研究试图进行评估和实验。 1）带有计时器的等间隔方法和2）剪切响应方法，重点关注场景的变化，从而通过这些方法自动准备摘要。剪切响应方法用于检测场景的变化评估结果显示，在切换前 2 秒取出声音（基于对节目的分析），并在切换后 3 秒录制场景。因此，可以提取剪辑、关键场景和关键词的变化，因此发现它的用处较少。