跨模态哈希学习研究进展  

A Survey on Cross-modal Hash Learning

作　　者：张博麟 陈征[1] ZHANG Bo-lin;CHEN Zheng(Faculty of Electrical Engineering and Computer Science,Ningbo University,Ningbo 315211,China)

机构地区：[1]宁波大学信息科学与工程学院,宁波315211

出　　处：《无线通信技术》2019年第4期35-39,44,共6页Wireless Communication Technology

摘　　要：相比于单模态的哈希算法,跨模态哈希更具价值和现实意义。跨模态哈希算法能够将不同模态的数据映射到共同的语义空间,从而实现不同模态数据之间的保相似性。随着生成对抗网络(GAN)近年来的发展,GAN已经显示出以无监督方式建模数据分布的能力。所以,本文将紧扣时代前沿,详细介绍跨模态哈希学习的研究现状以及基于生成对抗网络的跨模态哈希算法的最新研究成果。Compared to the single-modal hash algorithm,cross-modal hashing is more valuable and practical.The cross-modal hash algorithm can map different modal data to a common semantic space,thereby achieving similarity between different modal data.With the recent development of the Generation Against Network(GAN),GAN has shown the ability to model data distribution in an unsupervised manner.Therefore,this article will closely follow the frontiers of the times,detailing the research status of cross-modal hash learning and the latest research results of cross-modal hashing algorithms based on generating confrontation networks.

关 键 词：跨模态哈希学习 生成对抗网络 数据检索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]