Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

A Two-Stage Coarse-to-Fine Framework for Sparse Crowd Density Prediction in Digital Twin-Based Safety Monitoring.

Sensors (Basel, Switzerland)·2026
Same author

An Implementation of Inverse Cosine Hardware for Sound Rendering Applications.

Sensors (Basel, Switzerland)·2023
Same author

Effects of preoperative education using virtual reality on preoperative anxiety and information desire: a randomized clinical trial.

Journal of clinical monitoring and computing·2023
Same author

Predicting In Vivo Target Occupancy (TO) Profiles via Physiologically Based Pharmacokinetic-TO Modeling of Warfarin Pharmacokinetics in Blood: Importance of Low Dose Data and Prediction of Stereoselective Target Interactions.

Drug metabolism and disposition: the biological fate of chemicals·2023
Same author

Longitudinal association between adiposity changes and lung function deterioration.

Respiratory research·2023
Same author

Anatomical and Radiological Assessments of Injectate Spread Stratified by the Volume of the Pericapsular Nerve Group Block.

Anesthesia and analgesia·2023

相关实验视频

Updated: Jun 22, 2025

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

515

从单个图像中有效地去除雾,使用基于DCP的轻量级U-Net神经网络模型.

Yunho Han1, Jiyoung Kim1, Jinyoung Lee2

  • 1Department of Computer Science and Engineering, Sejong University, Seoul 05006, Republic of Korea.

Sensors (Basel, Switzerland)
|June 27, 2024
PubMed
概括

这项研究介绍了一种轻量级的U-net模型,用于有效地去除雾,显著提高了暗通道先行 (DCP) 方法的计算复杂性. 新型号在较少的参数下获得高质量的结果,使其适合于现实世界的应用.

关键词:
卷积神经网络是一种卷积神经网络.在黑暗道之前,黑暗道之前.在 Defog 中使用 Defog.德哈兹 (dehaze) 是一种消毒剂.图像恶化 图像恶化在U-net中,U-net是指U-net网络.

更多相关视频

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

2.7K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.0K

相关实验视频

Last Updated: Jun 22, 2025

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

515
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

2.7K
Deep Neural Networks for Image-Based Dietary Assessment
13:19

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

9.0K

科学领域:

  • 计算机视觉 计算机视觉
  • 图像处理 图像处理
  • 人工智能的人工智能

背景情况:

  • 传统的Dark Channel Prior (DCP) 雾去除方法存在很高的计算复杂性.
  • 这种复杂性阻碍了它们在实时或高分辨率图像和视频处理中的应用.
  • 由于计算需求,现有的方法很难在通用应用中实现.

研究的目的:

  • 提出一个轻量级的U-net神经网络模型,以有效地消除单图像雾.
  • 解决与传统的DCP算法相关的计算挑战.
  • 开发一种有效和高效的模型,用于实际的雾清除应用.

主要方法:

  • 提出了一种新的两级U-net架构,用加速卷曲取代复杂的DCP操作.
  • 该模型采用轻量化设计,参数数量少 (2百万),以实现高效的资源利用.
  • 该架构优化了从单个输入图像中快速有效地去除雾.

主要成果:

  • 拟议的模型实现了 26.65 dB 的平均峰值信号与噪声比 (PSNR) 和 0.88.8 的结构相似度指数 (SSIM).
  • 与传统的DCP相比,观察到显著的改善,平均PSNR增加11.5dB,SSIM增加0.22.
  • 该模型的性能与基于最先进的卷积神经网络 (CNN) 的方法相美,尽管它的尺寸更小.

结论:

  • 拟议的轻量级U-net模型为雾清除提供了有效和高效的解决方案.
  • 该模型成功地克服了传统DCP方法的计算限制.
  • 它的直观结构和高性能使其适用于各种资源有限的应用.