Jove
Visualize
Contact Us
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

Related Experiment Video

Updated: Oct 8, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

548

A Generative Adversarial Network Fused with Dual-Attention Mechanism and Its Application in Multitarget Image Fine

Jian Yin1, Zhibo Zhou2, Shaohua Xu1

  • 1College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao 266 590, China.

Computational Intelligence and Neuroscience
|December 28, 2021
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

Rapid Remodeling of Human White Adipose Tissue Following Bariatric Surgery.

bioRxiv : the preprint server for biology·2026
Same author

Midline congenital upper lip sinus: a rare clinical case with analytical review of diagnostic and therapy strategies.

Frontiers in pediatrics·2026
Same author

Ultra-performance Liquid Chromatography-mass Spectrometry-based metabolic profiling of prepubertal children with Turner syndrome.

Pediatric research·2026
Same author

Association between age-specific preconception thyroid-stimulating hormone (TSH) and birth weight: a retrospective study.

BMJ open·2026
Same author

Hydroxysafflor Yellow A Promotes Lac-Phe Synthesis to Suppress GIP and Ameliorate Obesity in DIO Mice.

Phytotherapy research : PTR·2025
Same author

Randomized controlled trial of Shikani optical stylet's one-person technique in nasotracheal intubation for children undergoing dental treatment.

Scientific reports·2025

A novel Attention-Mechanism Generative Adversarial Network (AM-GAN) enhances small-target detection and multitarget segmentation by improving feature clarity and boundary accuracy. This method excels in complex medical image segmentation tasks, even with limited data.

Area of Science:

  • Computer Vision
  • Medical Image Analysis
  • Artificial Intelligence

Background:

  • Challenges in multitarget complex image segmentation include insignificant morphological features, inaccurate small-target detection, unclear boundaries, and overlapping regions.
  • Existing methods struggle with small and unbalanced datasets, leading to suboptimal segmentation accuracy and detail loss.

Purpose of the Study:

  • To propose a generative adversarial network fused with an attention mechanism (AM-GAN) for improved multitarget complex image segmentation.
  • To enhance the detection of small targets, improve segmentation boundary clarity, and address issues of boundary overlap and offset deformation.

Main Methods:

  • Developed AM-GAN, integrating a generative network (residual network + nonlocal attention module) for feature extraction and enhancement.

More Related Videos

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

676

Related Experiment Videos

Last Updated: Oct 8, 2025

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

548
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

676
  • Utilized nonlocal spatial-channel dual attention to boost target features and segmentation boundary continuity.
  • Employed a fully convolutional network-based adversarial component to penalize information loss in small-target regions and improve detection accuracy.

    • Main Results:

    • AM-GAN demonstrated superior performance in medical MRI abdominal image segmentation (liver, kidney, spleen) on small, unbalanced datasets.
    • Achieved high accuracy metrics: 87.37% class pixel accuracy, 92.42% intersection over union, and 93% average Dice coefficient.
    • Significantly outperformed other methods in segmentation precision and accuracy, particularly for small targets and overlapping boundaries.

    Conclusions:

    • The proposed AM-GAN effectively addresses limitations in small-target detection and multitarget segmentation by leveraging attention mechanisms and GANs.
    • The method shows strong applicability for complex segmentation tasks, improving detail preservation and boundary definition.
    • AM-GAN offers a robust solution for medical image analysis and other segmentation challenges requiring high precision.