Satellite Image Inpainting With Edge-Conditional Expectation Attention

Satellite images often suffer from data loss and corruption due to various factors, including sensor malfunctions and atmospheric interference, leading to incomplete and degraded imagery. In satellite images, long-range dependencies are particularly significant due to irregular and widely distribute...

Full description

Saved in:
Bibliographic Details
Main Authors: Dazhi Zhou, Yanjun Chen, Yuhong Zhang, Jing Niu
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/10959709/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850281394826117120
author Dazhi Zhou
Yanjun Chen
Yuhong Zhang
Jing Niu
author_facet Dazhi Zhou
Yanjun Chen
Yuhong Zhang
Jing Niu
author_sort Dazhi Zhou
collection DOAJ
description Satellite images often suffer from data loss and corruption due to various factors, including sensor malfunctions and atmospheric interference, leading to incomplete and degraded imagery. In satellite images, long-range dependencies are particularly significant due to irregular and widely distributed geomorphological edges, such as rivers, mountains, and urban structures. Traditional convolutional neural network-based inpainting methods face challenges due to their fixed receptive fields and parameter sharing, limiting their ability to effectively capture long-range dependencies and differentiate between corrupted and uncorrupted areas. To address these limitations, we propose a deep learning approach based on an edge-conditional expectation attention module, which conditions the attention mechanism on edge information to enhance the model's focus on high-frequency edge details. This enables the network to capture critical structures within the image better. In addition, we apply Chebyshev’s inequality within the attention mechanism to constrain the expectation of attention outputs, reducing excessive deviations and stabilizing the reconstruction process. Experimental results demonstrate that our approach outperforms several state-of-the-art methods in restoring missing regions and reconnecting geomorphological features.
format Article
id doaj-art-de84d1529b2a4161ba6d6817ffa7d905
institution OA Journals
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-de84d1529b2a4161ba6d6817ffa7d9052025-08-20T01:48:20ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118108301084510.1109/JSTARS.2025.355920310959709Satellite Image Inpainting With Edge-Conditional Expectation AttentionDazhi Zhou0https://orcid.org/0009-0006-6951-2145Yanjun Chen1https://orcid.org/0000-0002-9111-6222Yuhong Zhang2Jing Niu3https://orcid.org/0000-0002-3932-5860School of Mathematical Sciences, Harbin Normal University, Harbin, ChinaSchool of Mathematical Sciences, Harbin Normal University, Harbin, ChinaSchool of Geographical Sciences, Harbin Normal University, Harbin, ChinaSchool of Mathematical Sciences, Harbin Normal University, Harbin, ChinaSatellite images often suffer from data loss and corruption due to various factors, including sensor malfunctions and atmospheric interference, leading to incomplete and degraded imagery. In satellite images, long-range dependencies are particularly significant due to irregular and widely distributed geomorphological edges, such as rivers, mountains, and urban structures. Traditional convolutional neural network-based inpainting methods face challenges due to their fixed receptive fields and parameter sharing, limiting their ability to effectively capture long-range dependencies and differentiate between corrupted and uncorrupted areas. To address these limitations, we propose a deep learning approach based on an edge-conditional expectation attention module, which conditions the attention mechanism on edge information to enhance the model's focus on high-frequency edge details. This enables the network to capture critical structures within the image better. In addition, we apply Chebyshev’s inequality within the attention mechanism to constrain the expectation of attention outputs, reducing excessive deviations and stabilizing the reconstruction process. Experimental results demonstrate that our approach outperforms several state-of-the-art methods in restoring missing regions and reconnecting geomorphological features.https://ieeexplore.ieee.org/document/10959709/Chebyshev adjustmentedge extractionimage inpaintingremote sensingtransformer
spellingShingle Dazhi Zhou
Yanjun Chen
Yuhong Zhang
Jing Niu
Satellite Image Inpainting With Edge-Conditional Expectation Attention
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Chebyshev adjustment
edge extraction
image inpainting
remote sensing
transformer
title Satellite Image Inpainting With Edge-Conditional Expectation Attention
title_full Satellite Image Inpainting With Edge-Conditional Expectation Attention
title_fullStr Satellite Image Inpainting With Edge-Conditional Expectation Attention
title_full_unstemmed Satellite Image Inpainting With Edge-Conditional Expectation Attention
title_short Satellite Image Inpainting With Edge-Conditional Expectation Attention
title_sort satellite image inpainting with edge conditional expectation attention
topic Chebyshev adjustment
edge extraction
image inpainting
remote sensing
transformer
url https://ieeexplore.ieee.org/document/10959709/
work_keys_str_mv AT dazhizhou satelliteimageinpaintingwithedgeconditionalexpectationattention
AT yanjunchen satelliteimageinpaintingwithedgeconditionalexpectationattention
AT yuhongzhang satelliteimageinpaintingwithedgeconditionalexpectationattention
AT jingniu satelliteimageinpaintingwithedgeconditionalexpectationattention