| 张晓宇.基于网络信息挖掘的视频博客自动语义标注[J].高技术通讯(中文),2013,23(1): |
| 基于网络信息挖掘的视频博客自动语义标注 |
| Automatic semantic annotation for video blogs based on web information mining |
| 修订日期：2012-03-10 |
| DOI： |
| 中文关键词: 信息挖掘， 语义标注， 标注扩展， 评价标准， 视频博客 |
| 英文关键词: information mining, semantic annotation, annotation expansion, evaluation criterion, video blog |
| 基金项目:中央级公益性科研院所基本科研业务费专项资金(XK2012-2, ZD2012-7-2)和中国科学技术信息研究所科研项目预研资金(YY201208)资助项目 |
|
| 摘要点击次数: 3089 |
| 全文下载次数: 2322 |
| 中文摘要: |
| 为获得高质量的视频博客语义标注，针对视频博客的特点，提出了一种基于网络信息挖掘的自动语义标注算法，该算法首先从分析视频博客自身所包含的信息入手，从中提取基本标注；然后借助丰富而便捷的网络资源，通过深入挖掘网络信息获取在底层特征和高层语义上都相关的信息，对基本标注进行改进和完善，最终实现基于上下文的标注扩展。为了更加全面客观地评价语义标注结果，提出了一种基于分值的评价标准，该标准有效兼顾了标注的正确性和完整性这两大重要指标，从而能更加准确地反映标注质量。实验结果表明，这种基于网络信息挖掘的自动语义标注算法能够显著提高语义标注质量，对于海量视频博客的有效获取与管理具有重要意义。 |
| 英文摘要: |
| This paper presents a new automatic algorithm for semantic annotation of video blogs based on web information mining with the aim to acquire high-quality annotation. The algorithm first extracts intrinsic annotations from the original content of the target video blogs, and then uses external resources to obtain informative contents that are relevant to the video blogs in both high level semantics and low level features and improve intrinsic annotations based on web information mining. Finally, the context-based annotation expansion is achieved. The paper also defines a new score-based evaluation criterion for multi-label annotation problems,which takes both the accuracy and the completeness in annotation into account. The experimental results demonstrate the effectiveness of the proposed annotation algorithm and evaluation criterion, both of which are significant for acquisition and management of a large amount of video blogs. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
