期刊文献+

基于概率图模型的多模态数据联合检索算法 被引量:1

Probabilistic graph model based joint retrieval algorithm for multi-modal data
在线阅读 下载PDF
导出
摘要 为了提高多模态检索的性能,提出一种多模态文档语义生成模型以及基于该模型的多模态数据联合检索算法。多模态文档语义生成模型认为文档中每个模态数据都是由相同的语义概念生成的,并且文档是多个模态数据的联合分布。为了简化模型的求解过程,假设各个模态数据之间的生成过程是相互独立的,于是可以对每个模态的条件概率进行单独计算。在多模态联合检索中,通过计算查询数据和待检索文档的联合概率来计算它们之间的相似度。实验结果表明,提出的方法与两步检索、语义索引和排序学习三种多模态检索方法相比具有更好的检索性能。此外,该方法可以扩展应用到具有三个及以上模态数据的文档联合检索中。 In order to improve the performance of the multi-modal retrieval, a semantic generation model of the multi-modal document and a joint retrieval algorithm for muhi-modal data based on this model are proposed. The semantic generation model of the multi-modal document deems that the data of each modal in the document is generated by the same semantic concept, and the documents complies with the joint distribution of data of multiple modals. In order to simplify the solving process of model, the conditional probability of eaeh modal can be calculated independently if assuming that the generation processes among all the modals' data are mutual independent. During the multi-modal joint retrieval, the joint probability of the query data and the document under retrieval are calculated to obtain the similarity between them. The experimental results show that this algorithm has better retrieval performance than the two-step retrieval algorithm, semantic indexing algorithm and ranking learning algo- rithm. In addition, the algorithm can be applied to the document joint retrieval of data with three and more modals.
出处 《现代电子技术》 北大核心 2017年第5期33-37,共5页 Modern Electronics Technique
基金 中国博士后科学基金面上项目(2014M560730) 四川省科技厅应用基础项目(2015JY0071) 成都师范学院高层次引进人才专项科研项目(YJRC2014-9) 成都师范学院自然科学类培育项目(CS14ZD02)
关键词 多模态检索 概率图模型 极大似然估计 产生式模型 multi-modal retrieval probabilistic graph model maximum likelihood estimation generative model
作者简介 甘胜江(1980-),重庆璧山人,硕士,讲师。研究方向为计算机应用。 孙连海(1974-),男,山西临县人,硕士,实验师。主要研究方向为计算机应用等。 何俊林(1977-),男,四川成都人,硕士。主要研究方向为软件工程。 卢颖(1978-),女,上海人,博士,副教授。主要研究方向为计算机网络等。
  • 相关文献

参考文献3

二级参考文献89

  • 1代科学,付畅俭,武德峰,李国辉.视频挖掘:概念、技术与应用[J].计算机应用研究,2006,23(1):1-4. 被引量:8
  • 2张静,路红,薛向阳.基于索引结构的高效运动视频检索[J].计算机研究与发展,2006,43(11):1953-1958. 被引量:3
  • 3庄毅,庄越挺,吴飞.Composite Distance Transformation for Indexing and κ-Nearest-Neighbor Searching in High-Dimensional Spaces[J].Journal of Computer Science & Technology,2007,22(2):208-217. 被引量:3
  • 4中国国家统计局2004年统计数据[OL].[2005-02-15].http://www.stats.gov.cn/tjsj/ndsj/2005/indexch.htm.
  • 5Peter Lyman, Varian Hal R. How much information [OL]. [2007-01-03]. http://www. sims. berkeley. edu/how muchinfo-2003
  • 6Liu Y N, Wu F. Video semantic concept detection using multi modality subspace correlation propagation [C]//Proc of the I3th Int Multimedia Modeling Conference. Berlin: Springer, 2007:527-534
  • 7Babaguchi N, Kawai Y, Kitahashi T. Event based indexing of broadcast sports video by intermodal collaboration [J]. IEEE Trans on Multimedia, 2002, 4(1): 68-75
  • 8Snoek C G M, Worring M. Multimedia event-based video indexing using time intervals [J]. IEEE Trans on Multimedia, 2005, 7(4): 638-647
  • 9Snoek C G M, Worring M, Smeulders A W M. Early versus late fusion in semantic video analysis [C] //Proc of the 13th Annual ACM Int Conf on Multimedia. New York: ACM, 2005 : 399-402
  • 10Hotellin H. The most predictable criterion [J]. Journal of Educational Psychology, 1935, 26:139-142

共引文献32

同被引文献7

引证文献1

二级引证文献15

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部