EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion

Salient object detection in remote sensing images (RSI-SOD) aims to identify visually prominent objects by mimicking human visual perception. While convolutional neural networks (CNNs) have significantly improved detection accuracy, most RSI-SOD methods suffer from high computational costs and large...

Full description

Saved in:
Bibliographic Details
Main Authors: Qian Tang, Zhen Wang, Xuqi Wang, Shan-Wen Zhang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10980003/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849325385662070784
author Qian Tang
Zhen Wang
Xuqi Wang
Shan-Wen Zhang
author_facet Qian Tang
Zhen Wang
Xuqi Wang
Shan-Wen Zhang
author_sort Qian Tang
collection DOAJ
description Salient object detection in remote sensing images (RSI-SOD) aims to identify visually prominent objects by mimicking human visual perception. While convolutional neural networks (CNNs) have significantly improved detection accuracy, most RSI-SOD methods suffer from high computational costs and large model sizes, limiting their applicability in resource-constrained environments. Additionally, RSI&#x2019;s complex backgrounds and diverse object scales further challenge existing methods. To address these issues, we propose EMHANet, a lightweight network that integrates edge texture detail extraction, multi-scale feature fusion, and hybrid attention mechanism. EMHANet consists of MobileNetV3 for feature extraction, an Edge Feature Integration Module (EFIM) for low-level edge details, a Multi-scale Contextual Information Enhancement Module (MCIEM) for high-level feature refinement, and a lightweight decoder for saliency prediction. The network employs a coarse-to-fine strategy to accurately detect salient objects while maintaining efficiency. Experiments on ORSSD and EORSSD datasets demonstrate EMHANet superior performance over 31 state-of-the-art methods. It achieves high accuracy with an inference speed of 143 fps, 0.257M parameters, and 0.92G FLOPs, making it suitable for resource-limited applications. The source code and dataset will be available on <uri>https://github.com/darkseid-arch/EMHANet</uri>
format Article
id doaj-art-084b1b0596dc443b944fc8f58d5ffbef
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-084b1b0596dc443b944fc8f58d5ffbef2025-08-20T03:48:26ZengIEEEIEEE Access2169-35362025-01-0113891648917810.1109/ACCESS.2025.356548510980003EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature FusionQian Tang0Zhen Wang1https://orcid.org/0000-0002-5765-0827Xuqi Wang2Shan-Wen Zhang3https://orcid.org/0000-0002-4534-5358School of Electronic Information, Xijing University, Xi&#x2019;an, ChinaSchool of Electronic Information, Xijing University, Xi&#x2019;an, ChinaSchool of Electronic Information, Xijing University, Xi&#x2019;an, ChinaSchool of Electronic Information, Xijing University, Xi&#x2019;an, ChinaSalient object detection in remote sensing images (RSI-SOD) aims to identify visually prominent objects by mimicking human visual perception. While convolutional neural networks (CNNs) have significantly improved detection accuracy, most RSI-SOD methods suffer from high computational costs and large model sizes, limiting their applicability in resource-constrained environments. Additionally, RSI&#x2019;s complex backgrounds and diverse object scales further challenge existing methods. To address these issues, we propose EMHANet, a lightweight network that integrates edge texture detail extraction, multi-scale feature fusion, and hybrid attention mechanism. EMHANet consists of MobileNetV3 for feature extraction, an Edge Feature Integration Module (EFIM) for low-level edge details, a Multi-scale Contextual Information Enhancement Module (MCIEM) for high-level feature refinement, and a lightweight decoder for saliency prediction. The network employs a coarse-to-fine strategy to accurately detect salient objects while maintaining efficiency. Experiments on ORSSD and EORSSD datasets demonstrate EMHANet superior performance over 31 state-of-the-art methods. It achieves high accuracy with an inference speed of 143 fps, 0.257M parameters, and 0.92G FLOPs, making it suitable for resource-limited applications. The source code and dataset will be available on <uri>https://github.com/darkseid-arch/EMHANet</uri>https://ieeexplore.ieee.org/document/10980003/Remote sensingsalient detectionedge featuresmulti-scale fusionattention mechanism
spellingShingle Qian Tang
Zhen Wang
Xuqi Wang
Shan-Wen Zhang
EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
IEEE Access
Remote sensing
salient detection
edge features
multi-scale fusion
attention mechanism
title EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
title_full EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
title_fullStr EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
title_full_unstemmed EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
title_short EMHANet: Lightweight Salient Object Detection for Remote Sensing Images via Edge-Aware Multiscale Feature Fusion
title_sort emhanet lightweight salient object detection for remote sensing images via edge aware multiscale feature fusion
topic Remote sensing
salient detection
edge features
multi-scale fusion
attention mechanism
url https://ieeexplore.ieee.org/document/10980003/
work_keys_str_mv AT qiantang emhanetlightweightsalientobjectdetectionforremotesensingimagesviaedgeawaremultiscalefeaturefusion
AT zhenwang emhanetlightweightsalientobjectdetectionforremotesensingimagesviaedgeawaremultiscalefeaturefusion
AT xuqiwang emhanetlightweightsalientobjectdetectionforremotesensingimagesviaedgeawaremultiscalefeaturefusion
AT shanwenzhang emhanetlightweightsalientobjectdetectionforremotesensingimagesviaedgeawaremultiscalefeaturefusion