QQ客服 官方微博 反馈留言

	Anchor Concept Graph Distance for Web Image Re-ranking
	Qiu Shi; Wang Xiaogang; Tang Xiaoou
	2013
会议名称	21st ACM International Conference on Multimedia, MM 2013
会议地点	Barcelona, Spain
英文摘要	Web image re-ranking aims to automatically refine the initial textbased image search results by employing visual information. A strong line of work in image re-ranking relies on building image graphs that requires computing distances between image pairs. In this paper, we present Anchor Concept Graph Distance (ACG Distance), a novel distance measure for image re-ranking. Fora given textual query, an Anchor Concept Graph (ACG) is automatically learned from the initial text-based search results. The nodes of the ACG (i.e., anchor concepts) and their correlations well model the semantic structure of the images to be re-ranked.Images are projected to the anchor concepts. The projection vectors undergo a diffusion process over the ACG, and then are used to compute the ACG distance. The ACG distance reduces the semantic gap and better represents distances betweenimages. Experiments on the MSRA-MM and INRIA datasets show that the ACG distance consistently outperforms existingdistance measures and significantly improves start-of-the-art methods in image re-ranking.
收录类别	EI
语种	英语
内容类型	会议论文
源URL	[http://ir.siat.ac.cn:8080/handle/172644/4497]
专题	深圳先进技术研究院_集成所
作者单位	2013
推荐引用方式 GB/T 7714	Qiu Shi,Wang Xiaogang,Tang Xiaoou. Anchor Concept Graph Distance for Web Image Re-ranking[C]. 见:21st ACM International Conference on Multimedia, MM 2013. Barcelona, Spain.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。