Deep Hashing Based on VAE-GAN for Efficient Similarity Retrieval

doi:10.1049/cje.2019.08.001

	Deep Hashing Based on VAE-GAN for Efficient Similarity Retrieval
	Jin, Guoqing 1,2; Zhang, Yongdong 1; Lu, Ke 2
	2019-11-01
发表期刊	CHINESE JOURNAL OF ELECTRONICS
ISSN	1022-4653
卷号	28 期号:6 页码:1191-1197
摘要	Inspired by the recent advances in generative networks, we propose a VAE-GAN based hashing framework for fast image retrieval. The method combines a Variational autoencoder (VAE) with a Generative adversarial network (GAN) to generate content preserving images for pairwise hashing learning. By accepting real image and systhesized image in a pairwise form, a semantic perserving feature mapping model is learned under a adversarial generative process. Each image feature vector in the pairwise is converted to a hash codes, which are used in a pairwise ranking loss that aims to preserve relative similarities on images. Extensive experiments on several benchmark datasets demonstrate that the proposed method shows substantial improvement over the state-of-the-art hashing methods.
关键词	file organisation image retrieval learning (artificial intelligence) neural nets pairwise hashing learning semantic perserving feature mapping model adversarial generative process image feature vector hash codes pairwise ranking loss generative networks VAE-GAN based hashing framework image retrieval content preserving images similarity retrieval variational autoencoder generative adversarial network Image retrieval Learning to hash Variational autoencoder(VAE) Generative adversarial network(GAN)
DOI	10.1049/cje.2019.08.001
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2017YFB1002203] ; National Nature Science Foundation of China[61525206] ; National Nature Science Foundation of China[61672495] ; National Nature Science Foundation of China[61771458] ; National Nature Science Foundation of China[61702479] ; National Nature Science Foundation of China[61571424]
WOS研究方向	Engineering
WOS类目	Engineering, Electrical & Electronic
WOS记录号	WOS:000524255900016
出版者	TECHNOLOGY EXCHANGE LIMITED HONG KONG
引用统计	被引频次：6[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/14128
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Zhang, Yongdong
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Jin, Guoqing,Zhang, Yongdong,Lu, Ke. Deep Hashing Based on VAE-GAN for Efficient Similarity Retrieval[J]. CHINESE JOURNAL OF ELECTRONICS,2019,28(6):1191-1197.
APA	Jin, Guoqing,Zhang, Yongdong,&Lu, Ke.(2019).Deep Hashing Based on VAE-GAN for Efficient Similarity Retrieval.CHINESE JOURNAL OF ELECTRONICS,28(6),1191-1197.
MLA	Jin, Guoqing,et al."Deep Hashing Based on VAE-GAN for Efficient Similarity Retrieval".CHINESE JOURNAL OF ELECTRONICS 28.6(2019):1191-1197.