EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multi-Feature Fusion Method Based on Adaptive Dilation Convolution for Small-Object Detection

Lin Cao iD, Jin Wu iD, Zongmin Zhao iD, Chong Fu iD, Dongfeng Wang iD

Sensors · 18.05.2025 · 25 (10) · 3182

DOI: 10.3390/s25103182 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 18.05.2025. Verfasst von Lin Cao und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Information and Communication Engineering, Beijing Information Science and Technology University, Beijing 100101, China. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Beijing Education Communication Project, National Science Foundation of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper addresses the challenge of small-object detection in traffic surveillance by proposing a hybrid network architecture that combines attention mechanisms with convolutional layers. The network introduces an innovative attention mechanism into the YOLOv8 backbone, which effectively enhances the detection accuracy and robustness of small objects through fine-grained and coarse-grained attention routing on feature maps. During the feature fusion stage, we employ adaptive dilated convolution, which dynamically adjusts the dilation rate spatially based on frequency components. This adaptive convolution kernel helps preserve the details of small objects while strengthening their feature representation. It also expands the receptive field, which is beneficial for capturing contextual information and the overall features of small objects. Our method demonstrates an improvement in Average Precision (AP) by 1% on the UA-DETRAC-test dataset and 3% on the VisDrone-test dataset when compared to state-of-the-art methods. The experiments indicate that the new architecture achieves significant performance improvements across various evaluation metrics. To fully leverage the potential of our approach, we conducted extended research on radar–camera systems.

Schlagwörter

Stichwörter der Autor:innen

adaptive dilated convolutionattention mechanismfeature fusionradar–camera systemsmall-object detection

Autor:innen und Institutionen

  • School of Information and Communication Engineering, Beijing Information Science and Technology University, Beijing 100101, China.
  • Key Laboratory of the Ministry of Education for Optoelectronic Measurement Technology and Instrument, Beijing Information Science and Technology University, Beijing 100101, China.
  • School of Computer Science and Engineering, Northeastern University, Shenyang 110169, China.
  • Beijing TransMicrowave Technology Company, Beijing 100080, China.

ORCID-Profile: Lin Cao, Jin Wu, Zongmin Zhao, Chong Fu, Dongfeng Wang

Förderung

  • Beijing Education Communication Project
  • National Science Foundation of China

Zitieren

APA

Cao, L., Wu, J., Zhao, Z., Fu, C., & Wang, D. (2025). Multi-Feature Fusion Method Based on Adaptive Dilation Convolution for Small-Object Detection. Sensors, 25(10), 3182. https://doi.org/10.3390/s25103182

Vancouver

Cao L, Wu J, Zhao Z, Fu C, Wang D. Multi-Feature Fusion Method Based on Adaptive Dilation Convolution for Small-Object Detection. Sensors. 2025;25(10):3182. doi:10.3390/s25103182

BibTeX
@article{Cao2025,
  author = {Cao, Lin and Wu, Jin and Zhao, Zongmin and Fu, Chong and Wang, Dongfeng},
  title = {Multi-Feature Fusion Method Based on Adaptive Dilation Convolution for Small-Object Detection},
  journal = {Sensors},
  year = {2025},
  volume = {25},
  number = {10},
  pages = {3182},
  doi = {10.3390/s25103182}
}
Kostenlos starten →