Image auto-annotation via tag-dependent random search over range-constrained visual neighbours

doi:10.1007/s11042-013-1811-3

	Image auto-annotation via tag-dependent random search over range-constrained visual neighbours
	Lin, Zijia 1; Ding, Guiguang 2; Hu, Mingqing 3
	2015-06-01
发表期刊	MULTIMEDIA TOOLS AND APPLICATIONS
ISSN	1380-7501
卷号	74 期号:11 页码:4091-4116
摘要	The quantity setting of visual neighbours can be critical for the performance of many previously proposed visual-neighbour-based (VNB) image auto-annotation methods. And in those methods, each candidate tag of a to-be-annotated image would be better to have its own trustworthy part of visual neighbours for score prediction. Hence in this paper we propose to use a constrained range rather than an identical and fixed number of visual neighbours for VNB methods to allow more flexible choices of neighbours, and then put forward a novel tag-dependent random search process to estimate the tag-dependent trust degrees of visual neighbours for each candidate tag. We further propose an effective image auto-annotation method termed TagSearcher based on a widely-used conditional probability model for auto-annotation, considering image-dependent weights of visual neighbours, tag-dependent trust degrees of visual neighbours and votes for a candidate tag from visual neighbours. Extensive experiments conducted on both a benchmark dataset and real-world web images present that the proposed TagSearcher can yield inspiring annotation performance and also reduce the performance sensitivity to the quantity setting of visual neighbours.
关键词	Image auto-annotation TagSearcher Tag-dependent random search Range-constrained visual neighbours
DOI	10.1007/s11042-013-1811-3
收录类别	SCI
语种	英语
资助项目	National Basic Research Project of China[2011CB707000] ; National Natural Science Foundation of China[61271394] ; National Natural Science Foundation of China[61005045]
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Information Systems ; Computer Science, Software Engineering ; Computer Science, Theory & Methods ; Engineering, Electrical & Electronic
WOS记录号	WOS:000354898800022
出版者	SPRINGER
引用统计	被引频次：4[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/9682
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Lin, Zijia
作者单位	1.Tsinghua Univ, Dept Comp Sci & Technol, Beijing 100084, Peoples R China 2.Tsinghua Univ, Sch Software, Beijing 100084, Peoples R China 3.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Lin, Zijia,Ding, Guiguang,Hu, Mingqing. Image auto-annotation via tag-dependent random search over range-constrained visual neighbours[J]. MULTIMEDIA TOOLS AND APPLICATIONS,2015,74(11):4091-4116.
APA	Lin, Zijia,Ding, Guiguang,&Hu, Mingqing.(2015).Image auto-annotation via tag-dependent random search over range-constrained visual neighbours.MULTIMEDIA TOOLS AND APPLICATIONS,74(11),4091-4116.
MLA	Lin, Zijia,et al."Image auto-annotation via tag-dependent random search over range-constrained visual neighbours".MULTIMEDIA TOOLS AND APPLICATIONS 74.11(2015):4091-4116.