MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection

Infrared small-target detection encounters significant challenges due to a low image signal-to-noise ratio, limited target size, and complex background noise. To address the issues of sparse feature loss for small targets during the down-sampling phase of the traditional U-Net network and the semant...

Full description

Saved in:
Bibliographic Details
Main Authors: Jingcui Ma, Nian Pan, Dengyu Yin, Di Wang, Jin Zhou
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/14/2502
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849732927922896896
author Jingcui Ma
Nian Pan
Dengyu Yin
Di Wang
Jin Zhou
author_facet Jingcui Ma
Nian Pan
Dengyu Yin
Di Wang
Jin Zhou
author_sort Jingcui Ma
collection DOAJ
description Infrared small-target detection encounters significant challenges due to a low image signal-to-noise ratio, limited target size, and complex background noise. To address the issues of sparse feature loss for small targets during the down-sampling phase of the traditional U-Net network and the semantic gap in the feature fusion process, a multilevel feature extraction and fusion attention network (MEFA-Net) is designed. Specifically, the dilated direction-sensitive convolution block (DDCB) is devised to collaboratively extract local detail features, contextual features, and Gaussian salient features via ordinary convolution, dilated convolution and parallel strip convolution. Furthermore, the encoder attention fusion module (EAF) is employed, where spatial and channel attention weights are generated using dual-path pooling to achieve the adaptive fusion of deep and shallow layer features. Lastly, an efficient up-sampling block (EUB) is constructed, integrating a hybrid up-sampling strategy with multi-scale dilated convolution to refine the localization of small targets. The experimental results confirm that the proposed algorithm model surpasses most existing recent methods. Compared with the baseline, the intersection over union (IoU) and probability of detection <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mrow><mi>P</mi></mrow><mrow><mi>d</mi></mrow></msub></mrow></semantics></math></inline-formula> of MEFA-Net on the IRSTD-1k dataset are increased by 2.25% and 3.05%, respectively, achieving better detection performance and a lower false alarm rate in complex scenarios.
format Article
id doaj-art-604720e877bf4747a2f5a13b6f4cd2b2
institution DOAJ
issn 2072-4292
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-604720e877bf4747a2f5a13b6f4cd2b22025-08-20T03:08:10ZengMDPI AGRemote Sensing2072-42922025-07-011714250210.3390/rs17142502MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target DetectionJingcui Ma0Nian Pan1Dengyu Yin2Di Wang3Jin Zhou4National Laboratory on Adaptive Optics, Chengdu 610209, ChinaNational Laboratory on Adaptive Optics, Chengdu 610209, ChinaAVIC Chengdu Aircraft Design & Research Institute, Chengdu 610091, ChinaAVIC Chengdu Aircraft Design & Research Institute, Chengdu 610091, ChinaNational Laboratory on Adaptive Optics, Chengdu 610209, ChinaInfrared small-target detection encounters significant challenges due to a low image signal-to-noise ratio, limited target size, and complex background noise. To address the issues of sparse feature loss for small targets during the down-sampling phase of the traditional U-Net network and the semantic gap in the feature fusion process, a multilevel feature extraction and fusion attention network (MEFA-Net) is designed. Specifically, the dilated direction-sensitive convolution block (DDCB) is devised to collaboratively extract local detail features, contextual features, and Gaussian salient features via ordinary convolution, dilated convolution and parallel strip convolution. Furthermore, the encoder attention fusion module (EAF) is employed, where spatial and channel attention weights are generated using dual-path pooling to achieve the adaptive fusion of deep and shallow layer features. Lastly, an efficient up-sampling block (EUB) is constructed, integrating a hybrid up-sampling strategy with multi-scale dilated convolution to refine the localization of small targets. The experimental results confirm that the proposed algorithm model surpasses most existing recent methods. Compared with the baseline, the intersection over union (IoU) and probability of detection <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><msub><mrow><mi>P</mi></mrow><mrow><mi>d</mi></mrow></msub></mrow></semantics></math></inline-formula> of MEFA-Net on the IRSTD-1k dataset are increased by 2.25% and 3.05%, respectively, achieving better detection performance and a lower false alarm rate in complex scenarios.https://www.mdpi.com/2072-4292/17/14/2502infrared small-target detectiondilated direction-sensitive convolutionencoder attention fusionefficient up-sampling
spellingShingle Jingcui Ma
Nian Pan
Dengyu Yin
Di Wang
Jin Zhou
MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
Remote Sensing
infrared small-target detection
dilated direction-sensitive convolution
encoder attention fusion
efficient up-sampling
title MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
title_full MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
title_fullStr MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
title_full_unstemmed MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
title_short MEFA-Net: Multilevel Feature Extraction and Fusion Attention Network for Infrared Small-Target Detection
title_sort mefa net multilevel feature extraction and fusion attention network for infrared small target detection
topic infrared small-target detection
dilated direction-sensitive convolution
encoder attention fusion
efficient up-sampling
url https://www.mdpi.com/2072-4292/17/14/2502
work_keys_str_mv AT jingcuima mefanetmultilevelfeatureextractionandfusionattentionnetworkforinfraredsmalltargetdetection
AT nianpan mefanetmultilevelfeatureextractionandfusionattentionnetworkforinfraredsmalltargetdetection
AT dengyuyin mefanetmultilevelfeatureextractionandfusionattentionnetworkforinfraredsmalltargetdetection
AT diwang mefanetmultilevelfeatureextractionandfusionattentionnetworkforinfraredsmalltargetdetection
AT jinzhou mefanetmultilevelfeatureextractionandfusionattentionnetworkforinfraredsmalltargetdetection