Scholarlink

<p>In the present paper, we elaborate an accelerated Stochastic Primal-Dual Fixed Point (SPDFP) approach to address the challenges posed by deconvolution problem using Nesterov’s momentum. Unlike traditional full gradient updates, SPDFP utilizes stochastic gradients, as discussed in prior literature, notably in&#xa0;(Zhu and Zhang. J Sci Comput 84(1):16&#xa0;<CitationRef CitationID="CR32">2020</CitationRef>). Through numerical experiments, we demonstrate the method’s enhanced efficiency compared to its predecessors, emphasizing its practical effectiveness. Our investigation goes beyond empirical validation, delving into theoretical foundations and examining convergence properties under specific assumptions. This analysis offers deeper insights into the method’s performance characteristics, shedding light on its applicability and limitations in real-world deconvolution scenarios.</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

An Accelerated Stochastic Primal-Dual Fixed Point Approach for Image Deblurring

  • 期刊论文

  • 发表日期：

    2025年3月17日

查看全文
  • Yasmine El Mobariki,
  • Amine Laghrib

摘要

In the present paper, we elaborate an accelerated Stochastic Primal-Dual Fixed Point (SPDFP) approach to address the challenges posed by deconvolution problem using Nesterov’s momentum. Unlike traditional full gradient updates, SPDFP utilizes stochastic gradients, as discussed in prior literature, notably in (Zhu and Zhang. J Sci Comput 84(1):16 2020). Through numerical experiments, we demonstrate the method’s enhanced efficiency compared to its predecessors, emphasizing its practical effectiveness. Our investigation goes beyond empirical validation, delving into theoretical foundations and examining convergence properties under specific assumptions. This analysis offers deeper insights into the method’s performance characteristics, shedding light on its applicability and limitations in real-world deconvolution scenarios.