EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

MEAC: A Multi-Scale Edge-Aware Convolution Module for Robust Infrared Small-Target Detection

Jinlong Hu iD, Tian Zhang, Ming Zhao iD

Sensors · 16.07.2025 · 25 (14) · 4442

DOI: 10.3390/s25144442 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 16.07.2025. Verfasst von Jinlong Hu und 2 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Seismology, China Earthquake Administration, Wuhan 430071, China. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Infrared small-target detection remains a critical challenge in military reconnaissance, environmental monitoring, forest-fire prevention, and search-and-rescue operations, owing to the targets’ extremely small size, sparse texture, low signal-to-noise ratio, and complex background interference. Traditional convolutional neural networks (CNNs) struggle to detect such weak, low-contrast objects due to their limited receptive fields and insufficient feature extraction capabilities. To overcome these limitations, we propose a Multi-Scale Edge-Aware Convolution (MEAC) module that enhances feature representation for small infrared targets without increasing parameter count or computational cost. Specifically, MEAC fuses (1) original local features, (2) multi-scale context captured via dilated convolutions, and (3) high-contrast edge cues derived from differential Gaussian filters. After fusing these branches, channel and spatial attention mechanisms are applied to adaptively emphasize critical regions, further improving feature discrimination. The MEAC module is fully compatible with standard convolutional layers and can be seamlessly embedded into various network architectures. Extensive experiments on three public infrared small-target datasets (SIRSTD-UAVB, IRSTDv1, and IRSTD-1K) demonstrate that networks augmented with MEAC significantly outperform baseline models using standard convolutions. When compared to eleven mainstream convolution modules (ACmix, AKConv, DRConv, DSConv, LSKConv, MixConv, PConv, ODConv, GConv, and Involution), our method consistently achieves the highest detection accuracy and robustness. Experiments conducted across multiple versions, including YOLOv10, YOLOv11, and YOLOv12, as well as various network levels, demonstrate that the MEAC module achieves stable improvements in performance metrics while slightly increasing computational and parameter complexity. These results validate the MEAC module’s significant advantages in enhancing the detection of small and weak objects and suppressing interference from complex backgrounds. These results validate MEAC’s effectiveness in enhancing weak small-target detection and suppressing complex background noise, highlighting its strong generalization ability and practical application potential.

Schlagwörter

Stichwörter der Autor:innen

Multi-Scale Edge-Aware Convolution (MEAC)attention mechanismsdifferential Gaussian edge extractionfeature fusioninfrared small-target detectionmulti-scale dilated convolution

Autor:innen und Institutionen

  • Institute of Seismology, China Earthquake Administration, Wuhan 430071, China.
  • Northwest Land and Resource Research Center, Shaanxi Normal University, Xi'an 710119, China.
  • School of Cybersecurity and Informatization, Wuxi University, Wuxi 214105, China.

ORCID-Profile: Jinlong Hu, Ming Zhao

Zitieren

APA

Hu, J., Zhang, T., & Zhao, M. (2025). MEAC: A Multi-Scale Edge-Aware Convolution Module for Robust Infrared Small-Target Detection. Sensors, 25(14), 4442. https://doi.org/10.3390/s25144442

Vancouver

Hu J, Zhang T, Zhao M. MEAC: A Multi-Scale Edge-Aware Convolution Module for Robust Infrared Small-Target Detection. Sensors. 2025;25(14):4442. doi:10.3390/s25144442

BibTeX
@article{Hu2025,
  author = {Hu, Jinlong and Zhang, Tian and Zhao, Ming},
  title = {MEAC: A Multi-Scale Edge-Aware Convolution Module for Robust Infrared Small-Target Detection},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {14},
  pages = {4442},
  doi = {10.3390/s25144442}
}
Kostenlos starten →