Show simple item record

dc.contributor.authorYu, Jian Songzh_CN
dc.contributor.authorCao, Dong Linzh_CN
dc.contributor.authorLi, Shao Zizh_CN
dc.contributor.authorLin, Da Zhenzh_CN
dc.contributor.author李绍滋zh_CN
dc.contributor.author林达真zh_CN
dc.contributor.author曹冬林zh_CN
dc.date.accessioned2015-07-22T02:39:53Z
dc.date.available2015-07-22T02:39:53Z
dc.date.issued2012zh_CN
dc.identifier.citationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2012:580-588zh_CN
dc.identifier.other20140917395016zh_CN
dc.identifier.urihttps://dspace.xmu.edu.cn/handle/2288/86841
dc.descriptionConference Name:2012 International Conference on Web Information Systems and Mining, WISM 2012. Conference Address: Chengdu, China. Time:October 26, 2012 - October 28, 2012.zh_CN
dc.descriptionXihua University; Leshan Normal Universityzh_CN
dc.description.abstractWe propose an Internet-search-based automatic image annotation feedback model, combining content-based and web-based image annotation, to solve the relevance assumption between the image and text and the limited volume of the database. In this model, we extract candidate labels from search results using web-based texts associated with the image, and then verify the final results by using Internet search results of candidate labels with content-based features. Experimental results show that this method can annotate the large-scale database with high accuracy, and achieve a 5.2% improvement on the basis of web-based automatic image annotation. 漏 Springer-Verlag Berlin Heidelberg 2012.zh_CN
dc.language.isoen_USzh_CN
dc.publisherSpringer Verlagzh_CN
dc.source.urihttp://dx.doi.org/10.1007/978-3-642-33469-6_72zh_CN
dc.subjectImage analysiszh_CN
dc.subjectInternetzh_CN
dc.subjectSearch engineszh_CN
dc.subjectWebsiteszh_CN
dc.titleA novel image annotation feedback model based on internet-searchzh_CN
dc.typeConferencezh_CN


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record