摘要
近年来,数字多媒体图像出现了爆炸式的增长,人们在互联网搜索过程中遇到的问题也越来越多,提高图像的搜索效率极具挑战性。图像搜索是图像领域的研究热点,目前已有多种搜索技术在各商业领域得到应用,但搜索的结果并不能完全满足用户的需求,"语义鸿沟"的存在使得搜索结果仍存在一定的噪声。图像重排序为解决此问题提供了很好的帮助,在初始搜索的基础上进行重排序可使搜索结果更加准确和丰富。文中着重介绍图像重排序技术的研究进展,对已有研究方法进行总结和分析,比较各自的优缺点以及近年来突破的主要关键技术;关注最新的研究进展,总结了目前图像重排序的典型数据集以及针对特定领域研究建立的数据集,并对图像重排序领域未来的发展进行了展望。
In recent years,with the rapid development of Internet technology and the popularity of multimedia terminals electronic products,improving the efficiency of image search is a challenge in the media retrieval.The research of image search is a hot issue in the field of image.At present,many current commercial image search techniques have been applied,but the search results cannot meet the needs of users because of the existence of"semantic gap".The search results still have some noise.Image search reordering is helpful to solve this problem.Based on the initial search,results can be more accurate and more abundant after reranking.In this paper,the research progress of image search reranking technology was introduced,and the current research methods were summarized and analyzed.The advantages and disadvantages of these methods and the key technologies in recent years were compared.The latest research progress and the future development of image search reranking and the future development were also given.
作者
赵小艳
刘宏哲
袁家政
杨少鹏
ZHAO Xiao-yan;LIU Hong-zhe;YUAN Jia-zheng;YANG Shao-peng(Beijing Key Laboratory of Information Service Engineering,Beijing Union University, Beijing 100101 ,China)
出处
《计算机科学》
CSCD
北大核心
2018年第5期15-23,共9页
Computer Science
基金
国家自然科学基金(61372148
61271369)
北京市自然科学基金(4152016)
国家科技支撑课题(2014BAK08B02)资助
关键词
图像搜索
重排序
聚类
分类
图
Image search
Reranking
Clustering
Classification
Map