MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection

Infrared small target detection (ISTD) technology has extensive applications in the military field. Due to the quality of imaging equipment and environmental interference, infrared small target images lack texture and structural information. Deep learning-based algorithms have achieved superior accu...

Full description

Saved in:
Bibliographic Details
Main Authors: Luping Zhang, Junhai Luo, Yian Huang, Fengyi Wu, Xingye Cui, Zhenming Peng
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/10770559/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850242739349749760
author Luping Zhang
Junhai Luo
Yian Huang
Fengyi Wu
Xingye Cui
Zhenming Peng
author_facet Luping Zhang
Junhai Luo
Yian Huang
Fengyi Wu
Xingye Cui
Zhenming Peng
author_sort Luping Zhang
collection DOAJ
description Infrared small target detection (ISTD) technology has extensive applications in the military field. Due to the quality of imaging equipment and environmental interference, infrared small target images lack texture and structural information. Deep learning-based algorithms have achieved superior accuracy in this field compared to traditional algorithms; however, these methods are often not designed with domain knowledge integration. In this article, we propose a multidirectional information-guided contextual network (MDIGCNet) for ISTD. The primary structure of this network adopts the U-Net architecture. To address the issue of lacking texture and structural information in the target images, we employ an integrated differential convolution (IDConv) module to extract richer image features during both the encoding and decoding stages. Skip connections in the network utilize a multidirectional gradient information extraction block (MGIEB) to obtain gradient features of infrared small targets. Our domain-inspired multidirectional Gaussian differential convolution (MGDC) module is employed to extract features of Gaussian-distributed small targets, enhancing the distinction between targets and backgrounds. Additionally, we designed a local-global feature fusion (LGFF) module incorporating an attention mechanism to merge shallow and deep features, thereby improving the efficiency of feature utilization within the model. Furthermore, since both IDConv and MGDC are parallel multiconvolutional kernel structures, reparameterization techniques are used to avoid excessive parameters and computational load. Experimental results on public datasets NUDT-SIRST, IRSTD-1k, and SIRST-Aug demonstrate that our algorithm outperforms other state-of-the-art methods in detection performance.
format Article
id doaj-art-b226debdb6534a49b5bf655f04ad7a2e
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-b226debdb6534a49b5bf655f04ad7a2e2025-08-20T02:00:13ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01182063207610.1109/JSTARS.2024.350825510770559MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target DetectionLuping Zhang0https://orcid.org/0000-0002-3990-3106Junhai Luo1https://orcid.org/0000-0002-8435-007XYian Huang2Fengyi Wu3https://orcid.org/0009-0005-7770-2363Xingye Cui4Zhenming Peng5https://orcid.org/0000-0002-4148-3331School of Information and Communication Engineering, Chengdu, ChinaSchool of Information and Communication Engineering, Chengdu, ChinaSchool of Information and Communication Engineering, Chengdu, ChinaSchool of Information and Communication Engineering, Chengdu, ChinaSchool of Information and Communication Engineering, Chengdu, ChinaSchool of Information and Communication Engineering, Chengdu, ChinaInfrared small target detection (ISTD) technology has extensive applications in the military field. Due to the quality of imaging equipment and environmental interference, infrared small target images lack texture and structural information. Deep learning-based algorithms have achieved superior accuracy in this field compared to traditional algorithms; however, these methods are often not designed with domain knowledge integration. In this article, we propose a multidirectional information-guided contextual network (MDIGCNet) for ISTD. The primary structure of this network adopts the U-Net architecture. To address the issue of lacking texture and structural information in the target images, we employ an integrated differential convolution (IDConv) module to extract richer image features during both the encoding and decoding stages. Skip connections in the network utilize a multidirectional gradient information extraction block (MGIEB) to obtain gradient features of infrared small targets. Our domain-inspired multidirectional Gaussian differential convolution (MGDC) module is employed to extract features of Gaussian-distributed small targets, enhancing the distinction between targets and backgrounds. Additionally, we designed a local-global feature fusion (LGFF) module incorporating an attention mechanism to merge shallow and deep features, thereby improving the efficiency of feature utilization within the model. Furthermore, since both IDConv and MGDC are parallel multiconvolutional kernel structures, reparameterization techniques are used to avoid excessive parameters and computational load. Experimental results on public datasets NUDT-SIRST, IRSTD-1k, and SIRST-Aug demonstrate that our algorithm outperforms other state-of-the-art methods in detection performance.https://ieeexplore.ieee.org/document/10770559/Difference convolutioninfrared small target detection (ISTD)multidirectional gradient information extractionreparameterization
spellingShingle Luping Zhang
Junhai Luo
Yian Huang
Fengyi Wu
Xingye Cui
Zhenming Peng
MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Difference convolution
infrared small target detection (ISTD)
multidirectional gradient information extraction
reparameterization
title MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
title_full MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
title_fullStr MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
title_full_unstemmed MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
title_short MDIGCNet: Multidirectional Information-Guided Contextual Network for Infrared Small Target Detection
title_sort mdigcnet multidirectional information guided contextual network for infrared small target detection
topic Difference convolution
infrared small target detection (ISTD)
multidirectional gradient information extraction
reparameterization
url https://ieeexplore.ieee.org/document/10770559/
work_keys_str_mv AT lupingzhang mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection
AT junhailuo mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection
AT yianhuang mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection
AT fengyiwu mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection
AT xingyecui mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection
AT zhenmingpeng mdigcnetmultidirectionalinformationguidedcontextualnetworkforinfraredsmalltargetdetection