Jove
Visualize
Contact Us

Related Experiment Video

Updated: Jun 6, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

A lightweight remote sensing small-object detection approach with scale-based dynamic loss and efficient multi-scale

Anying Xu1, Xuanyu Wu2,3, Wuzhong Yang1

  • 1School of Mechanical Engineering, Guizhou University, Guiyang, 550025, China.

Scientific Reports
|June 4, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Integration of radiomics, deep learning, transcriptomics, and metabolomics reveals prognostic risk stratification and underlying biological mechanisms in colorectal cancer.

NPJ precision oncology·2026
Same author

Periodic asymmetric field enhances electrofusion of nanoscale lipid systems.

Nature communications·2025
Same author

Global, regional, and national burden of tracheal, bronchus, and lung cancer attributable to polycyclic aromatic hydrocarbons occupational expose from 1990 to 2021: worsening or improving?

BMC public health·2025
Same author

Hemodynamics of active abdominal compression decompression CPR in a rat model of multiple rib fractures with asphyxial cardiac arrest.

Scientific reports·2025
Same author

Unveiling the global landscape of anabolic steroid residues in beef via integrated targeted quantification and high-resolution suspect screening.

Current research in food science·2025
Same author

Elucidating biomarkers of carcinogenesis in ulcerative colitis: insight from Mendelian randomization and transcriptomics.

Discover oncology·2025
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

This study introduces a lightweight deep learning model for accurate small-object detection in remote sensing. The approach enhances efficiency and precision for real-time applications on edge devices.

Area of Science:

  • Computer Vision
  • Remote Sensing
  • Machine Learning

Background:

  • Small-object detection in remote sensing is vital but challenging due to limited features, complex backgrounds, and localization errors.
  • Existing methods struggle with scale variation and occlusion, impacting accuracy in real-world scenarios.

Purpose of the Study:

  • To develop a lightweight and efficient approach for accurate small-object detection in remote sensing imagery.
  • To improve target-background discrimination and localization accuracy for enhanced automated processes.

Main Methods:

  • Proposed a lightweight backbone (A2C2fLite) and an Efficient Multi-scale Attention (EMA) module.
  • Integrated a Scale-based Dynamic (SD) loss function to handle scale variation and occlusion.
  • Evaluated the approach on a remote sensing dataset, comparing it against baseline YOLOv13.

Related Experiment Videos

Last Updated: Jun 6, 2026

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments
03:31

End-To-End Deep Neural Network for Salient Object Detection in Complex Environments

Published on: December 15, 2023

Main Results:

  • Achieved 93.6% precision, 86.6% recall, and 92.6% mAP with only 2.0M parameters and 5.7 GFLOPs.
  • Demonstrated a superior accuracy-efficiency trade-off, reducing parameters by 20% and GFLOPs by 8.1% compared to the baseline.
  • Visualization confirmed suppressed missed alarms and refined bounding-box localization.

Conclusions:

  • The proposed lightweight design offers practical advantages for real-time small-object detection on resource-constrained edge devices.
  • The approach provides an efficient, robust, and highly deployable solution for complex remote sensing applications.
  • This work advances automated processes in environmental monitoring and aerial surveillance.