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

Deconvolution01:20

Deconvolution

664
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...
664

You might also read

Related Articles

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

Sort by
Same author

Efficient Data Aggregation in Smart Grids: A Personalized Local Differential Privacy Scheme.

Sensors (Basel, Switzerland)·2026
Same author

Roles of miR-200 family members in lung cancer: more than tumor suppressors.

Future oncology (London, England)·2018
Same author

Alterations and structural resilience of the gut microbiota under dietary fat perturbations.

The Journal of nutritional biochemistry·2018
Same author

Development and Feasibility Testing of an mHealth (Text Message and WeChat) Intervention to Improve the Medication Adherence and Quality of Life of People Living with HIV in China: Pilot Randomized Controlled Trial.

JMIR mHealth and uHealth·2018
Same author

Meta-Analysis of Preclinical Studies of Fibrinolytic Therapy for Acute Lung Injury.

Frontiers in immunology·2018
Same author

Corrigendum to: The calcium sensor TaCBL4 and its interacting protein TaCIPK5 are required for wheat resistance to stripe rust fungus.

Journal of experimental botany·2018

Related Experiment Video

Updated: Mar 15, 2026

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment
07:12

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment

Published on: January 6, 2026

574

Driven by Deformable Convolution and Multi-Plane Scale Constraint: A Hazy Image Dehazing-Stitching System.

Sheng Hu1, Han Xiao1, Cong Liu1

  • 1School of Electrical and Electronic Engineering, Hubei University of Technology, Wuhan 430068, China.

Sensors (Basel, Switzerland)
|March 14, 2026
PubMed
Summary
This summary is machine-generated.

This study introduces a novel non-uniform dehazing method using Deformable Convolution v4 (DCNv4) and a Retinex-inspired Transformer to enhance image clarity for advanced driver assistance systems (ADASs) in fog. The approach improves perception and stitching accuracy in adverse weather.

Keywords:
deep learningdeformable convolutionfeature matchingimage dehazingimage stitching

More Related Videos

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects
10:16

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects

Published on: February 8, 2014

12.7K
Automated Joint Space Detection Improves Bone Segmentation Accuracy
06:45

Automated Joint Space Detection Improves Bone Segmentation Accuracy

Published on: November 28, 2025

253

Related Experiment Videos

Last Updated: Mar 15, 2026

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment
07:12

Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment

Published on: January 6, 2026

574
Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects
10:16

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects

Published on: February 8, 2014

12.7K
Automated Joint Space Detection Improves Bone Segmentation Accuracy
06:45

Automated Joint Space Detection Improves Bone Segmentation Accuracy

Published on: November 28, 2025

253

Area of Science:

  • Computer Vision
  • Artificial Intelligence
  • Autonomous Driving

Background:

  • Adverse weather, like fog, degrades image quality, impacting deep learning algorithms crucial for advanced driver assistance systems (ADASs).
  • Existing dehazing methods struggle with non-uniform fog, causing background information loss and poor image stitching in ADAS scenarios due to sensor differences and low-texture features.

Purpose of the Study:

  • To develop an advanced non-uniform dehazing and image stitching method for robust environmental perception in foggy conditions for autonomous driving.
  • To enhance image clarity, feature matching accuracy, and stitching quality in challenging ADAS scenarios.

Main Methods:

  • A Deformable Convolution v4 (DCNv4)-based transform-like network was designed for long-range dependence and adaptive spatial aggregation.
  • A lightweight Retinex-inspired Transformer was integrated for color correction and structure refinement.
  • A multi-plane scale constraint module using the LightGlue network and an adaptive fusion stitching method were employed to improve matching and stitching precision.

Main Results:

  • The proposed method significantly improved feature matching accuracy and homography matrix estimation precision.
  • Achieved superior Peak Signal-to-Noise Ratios (PSNRs) of 22.78 dB (NH-HAZE) and 24.34 dB (BRAS) compared to existing methods.
  • Demonstrated effective elimination of artifacts and transition zones in stitched images.

Conclusions:

  • The developed non-uniform dehazing and stitching technique offers a reliable environmental perception solution for autonomous driving in foggy conditions.
  • The method's effectiveness and practicality in improving image quality and perception for ADAS were verified.