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: Jan 16, 2026

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

1.0K

Contrastive Learning-Driven Image Dehazing with Multi-Scale Feature Fusion and Hybrid Attention Mechanism.

Huazhong Zhang1, Jiaozhuo Wang1, Xiaoguang Tu1,2

  • 1College of Aviation Electronic and Electrical Engineering, Civil Aviation Flight University of China, Chengdu 641450, China.

Journal of Imaging
|September 26, 2025
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Deconvolution01:20

Deconvolution

545
Deconvolution, also known as inverse filtering, is the process of extracting the impulse response from known input and output signals. This technique is vital in scenarios where the system's characteristics are unknown, and they must be inferred from the observable signals.
Deconvolution involves several mathematical techniques to derive the impulse response. One common approach is polynomial division. In this method, the input and output sequences are treated as coefficients of...
545

You might also read

Related Articles

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

Sort by
Same author

Improved CenterNet-Based Multimodal Object Detection for Low-Light and Complex Environments.

Sensors (Basel, Switzerland)·2026
Same author

Metabolomics‑driven, data‑augmented machine learning for predicting toxicity of microplastic mixtures.

Ecotoxicology and environmental safety·2026
Same author

Leveraging plasma concentrations to optimize extracorporeal treatment in acute diquat poisoning: a multi-center retrospective cohort study.

Clinical toxicology (Philadelphia, Pa.)·2025
Same author

AnomNet: A Dual-Stage Centroid Optimization Framework for Unsupervised Anomaly Detection.

Journal of imaging·2025
Same author

Machine learning-driven QSAR models for predicting the cytotoxicity of five common microplastics.

Toxicology·2024
Same author

Evaluation of staged autologous blood transfusion during extracorporeal membrane oxygenation decannulation: A retrospective study.

Heart & lung : the journal of critical care·2024

This study introduces a novel image dehazing method using contrastive learning and InfoNCE loss for improved robustness. The approach effectively preserves image details and outperforms existing methods in diverse, hazy conditions.

Area of Science:

  • Computer Vision
  • Image Processing

Background:

  • Image dehazing is crucial for visual enhancement but faces challenges with fine detail preservation and non-uniform degradation.
  • Existing methods struggle with robustness and adapting to diverse hazy scenes.

Purpose of the Study:

  • To develop a novel image dehazing method that enhances robustness and preserves fine details.
  • To address limitations in current dehazing techniques for complex visual scenes.

Main Methods:

  • A contrastive learning framework utilizing InfoNCE loss, treating hazy images as negative and clear images as positive samples.
  • Integration of multi-scale dynamic feature fusion with a hybrid attention mechanism.
  • Dynamically adjustable frequency band filters and refined attention modules for cross-scale detail capture.
Keywords:
InfoNCE loss functionhybrid attention mechanismimage dehazingmulti-scale dynamic feature fusion

Related Experiment Videos

Last Updated: Jan 16, 2026

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

1.0K

Main Results:

  • The proposed method demonstrates improved ability to distinguish haze artifacts from scene features.
  • Enhanced preservation of image structural integrity and fine-grained details.
  • Outperforms most existing methods on RESIDE-6K and RS-Haze datasets.

Conclusions:

  • The novel contrastive learning and feature fusion approach offers a robust solution for image dehazing.
  • The method shows significant potential for practical applications requiring high-quality visual enhancement.
  • Advances in attention mechanisms and feature fusion contribute to superior performance in challenging dehazing scenarios.