Unsupervised random walk manifold contrastive hashing for multimedia retrieval

Abstract With the rapid growth in both the variety and volume of data on networks, especially within social networks containing vast multimedia data such as text, images, and video, there is an urgent need for efficient methods to retrieve helpful information quickly. Due to their high computational...

Full description

Saved in:
Bibliographic Details
Main Authors: Yunfei Chen, Yitian Long, Zhan Yang, Jun Long
Format: Article
Language:English
Published: Springer 2025-02-01
Series:Complex & Intelligent Systems
Subjects:
Online Access:https://doi.org/10.1007/s40747-025-01814-y
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850065646872690688
author Yunfei Chen
Yitian Long
Zhan Yang
Jun Long
author_facet Yunfei Chen
Yitian Long
Zhan Yang
Jun Long
author_sort Yunfei Chen
collection DOAJ
description Abstract With the rapid growth in both the variety and volume of data on networks, especially within social networks containing vast multimedia data such as text, images, and video, there is an urgent need for efficient methods to retrieve helpful information quickly. Due to their high computational efficiency and low storage costs, unsupervised deep cross-modal hashing methods have become the primary method for managing large-scale multimedia data. However, existing unsupervised deep cross-modal hashing methods still need help with issues such as inaccurate measurement of semantic similarity information, complex network architectures, and incomplete constraints among multimedia data. To address these issues, we propose an Unsupervised Random Walk Manifold Contrastive Hashing (URWMCH) method, designing a simple deep learning architecture. First, we build a random walk-based manifold similarity matrix based on the random walk strategy and modal-individual similarity structure. Second, we construct intra- and inter-modal similarity preservation and coexistent similarity preservation loss based on contrastive learning to constrain the training of hash functions, ensuring that the hash codes contain complete semantic association information. Finally, we designed comprehensive experiments on the MIRFlickr-25K, NUS-WIDE, and MS COCO datasets to demonstrate the effectiveness and superiority of the proposed URWMCH method.
format Article
id doaj-art-cdc3e07b670a41f19e9a8be7a15d734d
institution DOAJ
issn 2199-4536
2198-6053
language English
publishDate 2025-02-01
publisher Springer
record_format Article
series Complex & Intelligent Systems
spelling doaj-art-cdc3e07b670a41f19e9a8be7a15d734d2025-08-20T02:48:57ZengSpringerComplex & Intelligent Systems2199-45362198-60532025-02-0111411410.1007/s40747-025-01814-yUnsupervised random walk manifold contrastive hashing for multimedia retrievalYunfei Chen0Yitian Long1Zhan Yang2Jun Long3Big Data Institute, School of Computer Science and Engineering, Central South UniversityData Science Institute, Vanderbilt UniversityBig Data Institute, School of Computer Science and Engineering, Central South UniversityBig Data Institute, School of Computer Science and Engineering, Central South UniversityAbstract With the rapid growth in both the variety and volume of data on networks, especially within social networks containing vast multimedia data such as text, images, and video, there is an urgent need for efficient methods to retrieve helpful information quickly. Due to their high computational efficiency and low storage costs, unsupervised deep cross-modal hashing methods have become the primary method for managing large-scale multimedia data. However, existing unsupervised deep cross-modal hashing methods still need help with issues such as inaccurate measurement of semantic similarity information, complex network architectures, and incomplete constraints among multimedia data. To address these issues, we propose an Unsupervised Random Walk Manifold Contrastive Hashing (URWMCH) method, designing a simple deep learning architecture. First, we build a random walk-based manifold similarity matrix based on the random walk strategy and modal-individual similarity structure. Second, we construct intra- and inter-modal similarity preservation and coexistent similarity preservation loss based on contrastive learning to constrain the training of hash functions, ensuring that the hash codes contain complete semantic association information. Finally, we designed comprehensive experiments on the MIRFlickr-25K, NUS-WIDE, and MS COCO datasets to demonstrate the effectiveness and superiority of the proposed URWMCH method.https://doi.org/10.1007/s40747-025-01814-yCross-modal hashingMultimedia dataManifold similarityIntra- and inter-modal
spellingShingle Yunfei Chen
Yitian Long
Zhan Yang
Jun Long
Unsupervised random walk manifold contrastive hashing for multimedia retrieval
Complex & Intelligent Systems
Cross-modal hashing
Multimedia data
Manifold similarity
Intra- and inter-modal
title Unsupervised random walk manifold contrastive hashing for multimedia retrieval
title_full Unsupervised random walk manifold contrastive hashing for multimedia retrieval
title_fullStr Unsupervised random walk manifold contrastive hashing for multimedia retrieval
title_full_unstemmed Unsupervised random walk manifold contrastive hashing for multimedia retrieval
title_short Unsupervised random walk manifold contrastive hashing for multimedia retrieval
title_sort unsupervised random walk manifold contrastive hashing for multimedia retrieval
topic Cross-modal hashing
Multimedia data
Manifold similarity
Intra- and inter-modal
url https://doi.org/10.1007/s40747-025-01814-y
work_keys_str_mv AT yunfeichen unsupervisedrandomwalkmanifoldcontrastivehashingformultimediaretrieval
AT yitianlong unsupervisedrandomwalkmanifoldcontrastivehashingformultimediaretrieval
AT zhanyang unsupervisedrandomwalkmanifoldcontrastivehashingformultimediaretrieval
AT junlong unsupervisedrandomwalkmanifoldcontrastivehashingformultimediaretrieval