Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization
Most of the existing cross-modal retrieval methods make use of labeled data to learn projection matrices for different modal data. These methods usually learn the original semantic space to bridge the heterogeneous gap, ignoring the rich semantic information contained in unlabeled data. Accordingly,...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8957531/ |
id |
doaj-2f2f3cc1a9424041a57fd32f70c6d2b7 |
---|---|
record_format |
Article |
spelling |
doaj-2f2f3cc1a9424041a57fd32f70c6d2b72021-03-30T03:08:56ZengIEEEIEEE Access2169-35362020-01-018142781428810.1109/ACCESS.2020.29662208957531Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph RegularizationGongwen Xu0https://orcid.org/0000-0003-4518-9911Xiaomei Li1Zhijun Zhang2Business School, Shandong Jianzhu University, Jinan, ChinaThe Second Hospital, Shandong University, Jinan, ChinaComputer Science and Technology School, Shandong Jianzhu University, Jinan, ChinaMost of the existing cross-modal retrieval methods make use of labeled data to learn projection matrices for different modal data. These methods usually learn the original semantic space to bridge the heterogeneous gap, ignoring the rich semantic information contained in unlabeled data. Accordingly, a semantic consistency cross-modal retrieval with semi-supervised graph regularization (SCCMR) algorithm is proposed, which integrates the prediction of labels and the optimization of projection matrices into a unified framework to ensure that the solution obtained is globally optimal. At the same time, the method uses graph embedding to consider the nearest neighbors in the potential subspace of paired images and texts as well as images and texts with the same semantics. l<sub>21</sub>-norm constraint is applied to the projection matrices to select the discriminative features for different modal data. The results show that our method outperforms several advanced methods on four commonly used cross-modal retrieval datasets.https://ieeexplore.ieee.org/document/8957531/Cross-modal retrievalsemi-supervisedgraph regularizationsubspace learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Gongwen Xu Xiaomei Li Zhijun Zhang |
spellingShingle |
Gongwen Xu Xiaomei Li Zhijun Zhang Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization IEEE Access Cross-modal retrieval semi-supervised graph regularization subspace learning |
author_facet |
Gongwen Xu Xiaomei Li Zhijun Zhang |
author_sort |
Gongwen Xu |
title |
Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization |
title_short |
Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization |
title_full |
Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization |
title_fullStr |
Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization |
title_full_unstemmed |
Semantic Consistency Cross-Modal Retrieval With Semi-Supervised Graph Regularization |
title_sort |
semantic consistency cross-modal retrieval with semi-supervised graph regularization |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
Most of the existing cross-modal retrieval methods make use of labeled data to learn projection matrices for different modal data. These methods usually learn the original semantic space to bridge the heterogeneous gap, ignoring the rich semantic information contained in unlabeled data. Accordingly, a semantic consistency cross-modal retrieval with semi-supervised graph regularization (SCCMR) algorithm is proposed, which integrates the prediction of labels and the optimization of projection matrices into a unified framework to ensure that the solution obtained is globally optimal. At the same time, the method uses graph embedding to consider the nearest neighbors in the potential subspace of paired images and texts as well as images and texts with the same semantics. l<sub>21</sub>-norm constraint is applied to the projection matrices to select the discriminative features for different modal data. The results show that our method outperforms several advanced methods on four commonly used cross-modal retrieval datasets. |
topic |
Cross-modal retrieval semi-supervised graph regularization subspace learning |
url |
https://ieeexplore.ieee.org/document/8957531/ |
work_keys_str_mv |
AT gongwenxu semanticconsistencycrossmodalretrievalwithsemisupervisedgraphregularization AT xiaomeili semanticconsistencycrossmodalretrievalwithsemisupervisedgraphregularization AT zhijunzhang semanticconsistencycrossmodalretrievalwithsemisupervisedgraphregularization |
_version_ |
1724183959546363904 |