摘要
为了提高图像检索中的重排序效果,提出了一种基于图的多模态随机游走重排序算法。不同于现有的重排序算法根据检索返回的图像顺序设置图像列表得分序列初值,该算法将多模态融合应用于随机游走算法,避免单一模态获取图像内容的片面性,并利用多模态随机游走方法对返回图像列表得分序列进行初始化,然后利用多模态重排序算法最优化目标函数,对相关参数和得分列表进行迭代更新,从而获得最终重排序后的图像序列。实验显示了所提出的算法具有良好的重排序效果。
To improve the effect of re-ranking algorithms in image retrieval,this paper presented a multimodal graph- based re-ranking through random walk. Different from existing reranking algorithms which set the initial score se-quence value of an image list according to the image sequence returned by retrieval,the proposed method integrated multimodal to acquire more information and employed a multimodal random walk algorithm to initialize the relevance score list of the retrieved images. T h e n,the proposed method optimized the objective function by using a multimodal graph-based reranking algorithm in which an iteration procedure was used to update the parameters and relevance score list . Finally,the retrieved images were reordered according to the relevance score list . Experimental results demonstrate that the proposed reranking algorithm performs better than some other state- of- the-art algorithms.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2016年第10期1387-1393,共7页
Journal of Harbin Engineering University
基金
国家自然科学基金项目(61602004
61472001)
安徽省自然科学基金项目(1408085MF122
1508085MF127)
安徽省高校自然科学研究重点项目(KJ2016A041)
安徽大学信息保障技术协同创新中心公开招标课题(ADXXBZ2014-5ADXXBZ2014-6)
关键词
图像检索
多模态
随机游走
重排序
基于图的学习
image retrieval
multimodal
random walk
re- ranking
graph-based learning