首页> 外文会议>2013 Second International Conference on e-Learning and e-Technologies in Education >Tree-structure outline generation for lecture videos with synchronized slides
【24h】

Tree-structure outline generation for lecture videos with synchronized slides

机译:带有同步幻灯片的演讲视频的树形结构轮廓生成

获取原文
获取原文并翻译 | 示例

摘要

In this paper we propose a solution to generate tree-structure outline for lecture videos by analyzing their synchronized slides, by which detailed lecture overview can be automatically provided to E-learning portal users. Starting with OCR (Optical Character Recognition) result, we reconstruct the content of each slide. After that, we explore logical relations between slides, in order to make them hierarchical. And all potential redundant content will also be removed. Our evaluation shows that, based on our test dataset, the final outline achieved retains about 1/4 of the original texts from all slides and is organized well in an up-to-6-level tree structure. Furthermore, the average accuracy of all slide titles, which are undoubtedly the most important information, reaches 86%.
机译:在本文中,我们提出了一种解决方案,可以通过分析演讲视频的同步幻灯片来生成演讲视频的树形结构轮廓,从而可以将详细的演讲概述自动提供给电子学习门户用户。从OCR(光学字符识别)结果开始,我们重建每张幻灯片的内容。之后,我们探索幻灯片之间的逻辑关系,以使它们分层。并且所有潜在的冗余内容也将被删除。我们的评估表明,根据我们的测试数据集,最终的轮廓保留了所有幻灯片的原始文本的约1/4,并以多达6级的树状结构组织得很好。此外,无疑是最重要的信息,所有幻灯片标题的平均准确度达到86%。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号