期刊
NEUROCOMPUTING
卷 492, 期 -, 页码 264-277出版社
ELSEVIER
DOI: 10.1016/j.neucom.2022.04.011
关键词
Cross-modal retrieval; Matrix factorization; Hashing; Semantic information
资金
- Natural Science Foundation of China [71772107, 62072288]
- Natural Science Foundation of Shandong Province of China [ZR2020MF044, ZR202102230289, ZR2019MF003, ZR2021MF104]
- Shandong Education Quality Improvement Plan for Postgraduate (2021)
- SDUST Research Fund
- Humanity and Social Science Fund of the Ministry of Education [20YJAZH078, 20YJAZH127]
The proposed Dual Semantic Preserving Hashing (DSPH) method addresses the challenges of semantic information utilization and discriminative hash code learning in cross-modal hashing by leveraging matrix factorization and discrete optimization strategy.
Hashing methods have recently received widespread attention due to their flexibility and effectiveness for cross-modal retrieval tasks. However, most existing cross-modal hashing methods have some chal-lenging problems, in particular, effective exploitation of semantic information and learning discrimina-tive hash codes. To address these challenges, we propose an efficient Dual Semantic Preserving Hashing (DSPH) method, which first leverages matrix factorization to obtain low-level latent semantic representations of different modalities and remove redundant information. To enhance the discrimina-tive capability of hash codes, we preserve the high-level pairwise semantics and the learned low-level latent semantics into the unified hash codes. Finally, DSPH adopts discrete optimization strategy to learn the hash codes directly. Experimental results on three benchmark datasets demonstrate that the pro-posed DSPH method outperforms many state-of-the-art cross-modal hashing methods in terms of retrie-val accuracy, especially when dealing with short hash code. (c) 2022 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据