Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning

Remote sensing image captioning aims to generate semantically accurate descriptions that are closely linked to the visual features of remote sensing images. Existing approaches typically emphasize fine-grained extraction of visual features and capturing global information. However, they often overlo...

Full description

Saved in:
Bibliographic Details
Main Authors: Maofu Liu, Jiahui Liu, Xiaokang Zhang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11039674/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849422298965082112
author Maofu Liu
Jiahui Liu
Xiaokang Zhang
author_facet Maofu Liu
Jiahui Liu
Xiaokang Zhang
author_sort Maofu Liu
collection DOAJ
description Remote sensing image captioning aims to generate semantically accurate descriptions that are closely linked to the visual features of remote sensing images. Existing approaches typically emphasize fine-grained extraction of visual features and capturing global information. However, they often overlook the complementary role of textual information in enhancing visual semantics and face challenges in precisely locating objects that are most relevant to the image context. To address these challenges, this article presents a semantic–spatial feature fusion with dynamic graph refinement (SFDR) method, which integrates the semantic–spatial feature fusion (SSFF) and dynamic graph feature refinement (DGFR) modules. The SSFF module utilizes a multilevel feature representation strategy by leveraging pretrained CLIP features, grid features, and ROI features to integrate rich semantic and spatial information. In the DGFR module, a graph attention network captures the relationships between feature nodes, while a dynamic weighting mechanism prioritizes objects that are most relevant to the current scene and suppresses less significant ones. Therefore, the proposed SFDR method significantly enhances the quality of the generated descriptions. Experimental results on three benchmark datasets demonstrate the effectiveness of the proposed method.
format Article
id doaj-art-e0754bcfb3d64c7ba19db2ccf1ce9424
institution Kabale University
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-e0754bcfb3d64c7ba19db2ccf1ce94242025-08-20T03:31:10ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118154421545510.1109/JSTARS.2025.358068611039674Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image CaptioningMaofu Liu0https://orcid.org/0000-0002-3732-4354Jiahui Liu1https://orcid.org/0009-0008-9693-0734Xiaokang Zhang2https://orcid.org/0000-0002-6127-4801School of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, ChinaSchool of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, ChinaSchool of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan, ChinaRemote sensing image captioning aims to generate semantically accurate descriptions that are closely linked to the visual features of remote sensing images. Existing approaches typically emphasize fine-grained extraction of visual features and capturing global information. However, they often overlook the complementary role of textual information in enhancing visual semantics and face challenges in precisely locating objects that are most relevant to the image context. To address these challenges, this article presents a semantic–spatial feature fusion with dynamic graph refinement (SFDR) method, which integrates the semantic–spatial feature fusion (SSFF) and dynamic graph feature refinement (DGFR) modules. The SSFF module utilizes a multilevel feature representation strategy by leveraging pretrained CLIP features, grid features, and ROI features to integrate rich semantic and spatial information. In the DGFR module, a graph attention network captures the relationships between feature nodes, while a dynamic weighting mechanism prioritizes objects that are most relevant to the current scene and suppresses less significant ones. Therefore, the proposed SFDR method significantly enhances the quality of the generated descriptions. Experimental results on three benchmark datasets demonstrate the effectiveness of the proposed method.https://ieeexplore.ieee.org/document/11039674/Dynamic weighting mechanismfeature fusiongraph attention (GAT)remote sensing image captioning
spellingShingle Maofu Liu
Jiahui Liu
Xiaokang Zhang
Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Dynamic weighting mechanism
feature fusion
graph attention (GAT)
remote sensing image captioning
title Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
title_full Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
title_fullStr Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
title_full_unstemmed Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
title_short Semantic–Spatial Feature Fusion With Dynamic Graph Refinement for Remote Sensing Image Captioning
title_sort semantic x2013 spatial feature fusion with dynamic graph refinement for remote sensing image captioning
topic Dynamic weighting mechanism
feature fusion
graph attention (GAT)
remote sensing image captioning
url https://ieeexplore.ieee.org/document/11039674/
work_keys_str_mv AT maofuliu semanticx2013spatialfeaturefusionwithdynamicgraphrefinementforremotesensingimagecaptioning
AT jiahuiliu semanticx2013spatialfeaturefusionwithdynamicgraphrefinementforremotesensingimagecaptioning
AT xiaokangzhang semanticx2013spatialfeaturefusionwithdynamicgraphrefinementforremotesensingimagecaptioning