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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Enhancing anomaly-based zero-day attack detection framework using CNN-driven feature extraction and OC-SVM.

Scientific reports·2026
Same author

Hybrid GA-DQL approach for efficient task mapping of IoT applications in fog computing framework.

Scientific reports·2026
Same author

An ensemble-based sentiment analysis approach for precision medicine recommendation.

Scientific reports·2026
Same author

Energy and makespan optimised task mapping in fog enabled IoT application: a hybrid approach.

Scientific reports·2026
Same author

Cloud-enabled automatic modulation classification using deep feature fusion and Moth-Flame Optimized ELM approach.

Scientific reports·2025
Same author

Evolutionary optimization-based descendent adaptive filter for noise confiscation in electrocardiogram signals.

Physical and engineering sciences in medicine·2025

Related Experiment Video

Updated: Sep 17, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.0K

An efficient camouflaged image segmentation with modified UNet and attention techniques.

Isha Padhy1,2, Prabhat Dansena1, Sampa Sahoo1

  • 1Department of Computer Science, C V Raman Global University, Bhubaneswar, India.

Scientific Reports
|July 2, 2025
PubMed
Summary

Camouflaged object segmentation (COS) is improved by CAMO-UNet, a novel model using residual blocks and attention mechanisms. This computer vision approach enhances feature learning and accuracy for challenging segmentation tasks.

Keywords:
Attention mechanismCamouflage object segmentationComputer visionDeep learningObject detectionUNet

More Related Videos

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.6K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

529

Related Experiment Videos

Last Updated: Sep 17, 2025

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.0K
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
05:56

Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application

Published on: April 14, 2023

2.6K
Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique
04:48

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

529

Area of Science:

  • Computer Vision
  • Image Processing
  • Artificial Intelligence

Background:

  • Camouflaged object segmentation (COS) is a complex computer vision problem due to objects blending with backgrounds.
  • Standard models like UNet face limitations with ambiguous boundaries and texture similarity, leading to segmentation errors.
  • Existing methods struggle with precise object separation in challenging camouflage scenarios.

Purpose of the Study:

  • To introduce CAMO-UNet, a novel deep learning architecture designed to overcome the limitations of traditional models in camouflaged object segmentation.
  • To enhance the accuracy and robustness of segmentation models for objects with intricate camouflage patterns.
  • To improve feature representation and learning efficiency in deep neural networks for computer vision tasks.

Main Methods:

  • The proposed CAMO-UNet integrates residual blocks to facilitate gradient flow and enable deeper network architectures.
  • An attention mechanism is incorporated to focus on salient features and capture long-range spatial dependencies.
  • A Depth-aware triangular cyclic learning rate (CLR) strategy is employed to optimize training dynamics across different network depths.

Main Results:

  • CAMO-UNet achieved a high accuracy of 93.8% on benchmark datasets for camouflaged object segmentation.
  • The model demonstrated superior performance compared to state-of-the-art methods, including SINet, BGNet, and PFNet.
  • Evaluations showed significant improvements in key segmentation metrics such as S-measure, F-measure, and Mean Absolute Error (MAE).

Conclusions:

  • CAMO-UNet effectively addresses the challenges of camouflaged object segmentation through its innovative architectural components.
  • The integration of residual blocks, attention mechanisms, and CLR contributes to enhanced feature learning and segmentation accuracy.
  • The model represents a significant advancement in computer vision for accurately segmenting camouflaged objects in complex environments.