Scholarlink

<p>Image deblurring is a highly challenging task that primarily involves accurately estimating blur kernels and recovering clear, detailed images from given blurry images. To address this challenge, numerous image prior algorithms have been extensively explored and proposed. Given the significant impact of nonlinear channel priors in recent deblurring research, this paper introduces an enhanced rational channel prior (ERC) algorithm specifically designed for blind image deblurring. The motivation behind ERC is to enhance the ratio between the dark channel and enhanced local maximum intensity channel prior, which tends to be affected by blurring operations. Experimental results demonstrate that ERC exhibits superior discriminative capability compared to nonlinear channel priors. The model integrates the <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="34_2025_3211_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="19" /> </InlineMediaObject> <EquationSource Format="TEX">\({{L}_{1}}\)</EquationSource> </InlineEquation> norm into the ERC component and embeds it within the conventional deblurring architecture. Building on this, We introduce a new energy function and adopt an innovative optimization approach that integrates the half-quadratic splitting technique with the fast iterative shrinkage-thresholding algorithm for solving this energy function. Extensive experiments show that our method outperforms current mainstream approaches in terms of effectiveness.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Blind Image Deblurring via Enhanced Rational Channel Prior

  • 期刊论文

  • 发表日期：

    2025年7月9日

查看全文
  • Naimang Hu,
  • Jieqing Tan,
  • Xianyu Ge,
  • Dandan Hu,
  • Jing Liu

摘要

Image deblurring is a highly challenging task that primarily involves accurately estimating blur kernels and recovering clear, detailed images from given blurry images. To address this challenge, numerous image prior algorithms have been extensively explored and proposed. Given the significant impact of nonlinear channel priors in recent deblurring research, this paper introduces an enhanced rational channel prior (ERC) algorithm specifically designed for blind image deblurring. The motivation behind ERC is to enhance the ratio between the dark channel and enhanced local maximum intensity channel prior, which tends to be affected by blurring operations. Experimental results demonstrate that ERC exhibits superior discriminative capability compared to nonlinear channel priors. The model integrates the \({{L}_{1}}\) norm into the ERC component and embeds it within the conventional deblurring architecture. Building on this, We introduce a new energy function and adopt an innovative optimization approach that integrates the half-quadratic splitting technique with the fast iterative shrinkage-thresholding algorithm for solving this energy function. Extensive experiments show that our method outperforms current mainstream approaches in terms of effectiveness.