[1]徐明亮,余肖生.基于跨模态检索的效率优化算法[J].计算机技术与发展,2019,29(11):67-70.[doi:10. 3969 / j. issn. 1673-629X. 2019. 11. 014]
　XU Ming-liang,YU Xiao-sheng.An Efficiency Optimization Algorithm Based on Cross-modal Retrieval[J].,2019,29(11):67-70.[doi:10. 3969 / j. issn. 1673-629X. 2019. 11. 014]
点击复制

基于跨模态检索的效率优化算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
29
期数:
2019年11期
页码:
67-70
栏目:
智能、算法、系统工程
出版日期:
2019-11-10

文章信息/Info

Title:
An Efficiency Optimization Algorithm Based on Cross-modal Retrieval
文章编号:
1673-629X(2019)11-0067-04
作者:
徐明亮; 余肖生
三峡大学 计算机与信息学院,湖北 宜昌 443002
Author(s):
XU Ming-liang; YU Xiao-sheng
School of Computer and Information,Three Gorges University,Yichang 443002,China
关键词:
跨模态检索; 语义鸿沟; 典型相关分析; 主成分分析; 子空间投影
Keywords:
cross-modal search; semantic gap; canonical correlation analysis; principal component analysis; subspace projection
分类号:
TP301
DOI:
10. 3969 / j. issn. 1673-629X. 2019. 11. 014
摘要:
人们对于信息的需求已从单一的文本发展到图片、视频、声音等多种类型。 利用跨模态检索从不同类型的数据中同时找到表示同一信息的数据,已经成为满足人们信息需求的有效途径,也成为了信息检索领域的研究热点。 传统的跨模态检索算法由于采用的是经典的典型分析法,该算法存在一定的局限性和缺陷。 为了提高传统算法的检索效率,针对传统跨模态检索算法在处理高维度计算量巨大的问题,提出了一种跨模态信息检索的优化方法。 将传统的跨模态检索算法与主成分分析法相结合,提出一种新的信息检索算法,并进行了相应的实验测试。 实验结果表明,与传统算法相比,该方法在保证查准率基本不变的情况下,可以大幅减少原有算法的计算量,提高检索效率。
Abstract:
People’s demand for information has evolved from a single text to pictures,video,sound and other types. Using cross-modal retrieval to find data representing the same information from different types of data has become an effective way to meet people’s information needs and a research hotspot in the field of information retrieval. The traditional cross-modal retrieval algorithm is based on the classical typical analysis method,which has certain limitations and defects. In order to improve the retrieval efficiency of traditional algorithms,we propose an optimization method for cross - modal information retrieval to deal with the problem of large-scale computational complexity in the traditional cross-modal retrieval algorithm. Combining the traditional cross-modal retrieval algorithm with principal component analysis,a new information retrieval algorithm is proposed,and the corresponding experiment is carried out for testing. The experiment shows that compared with the traditional algorithm,the proposed method can greatly reduce the calculation of the original algorithm and improve the retrieval efficiency while ensuring that the precision is basically unchanged.

相似文献/References:

[1]乔荣华 周明全 耿国华.基于语义分类的文物图像标注研究[J].计算机技术与发展,2007,(07):200.
　QIAO Rong-hua,ZHOU Ming-quan,GENG Guo-hua.Research on Semantic Annotation to Cultural Relic Images Based on Semantic Classification[J].,2007,(11):200.
[2]邵曦,郁青玲. 基于典型相关的音乐跨模态检索[J].计算机技术与发展,2015,25(07):76.
　SHAO Xi,YU Qing-ling. Cross-modal Music Retrieval Based on Canonical Correlation[J].,2015,25(11):76.
[3]李新卫,吴飞,荆晓远.基于协同矩阵分解的单标签跨模态检索[J].计算机技术与发展,2018,28(11):99.[doi:10.3969/ j. issn.1673-629X.2018.11.022]
　LI Xin-wei,WU Fei,JING Xiao-yuan.Cross-modality Retrieval Based on Collective Matrix Factorization with Single Label[J].,2018,28(11):99.[doi:10.3969/ j. issn.1673-629X.2018.11.022]

更新日期/Last Update: 2019-11-10