Scholarlink

<p>Low-light photography often results in degraded images with poor visibility and motion blur. Conventional methods typically address enhancement and deblurring sequentially, which can accumulate errors and yield suboptimal restoration results. We propose CPM-LED, an end-to-end network featuring a Curve Complementary Attention module for balanced enhancement, a Parallel Hybrid Attention module for deep feature representation, and a Multi-Scale Feature Enhancement module for effective deblurring. Experimental results demonstrate competitive performance with a PSNR of 27.481, SSIM of 0.892, and LPIPS of 0.202, highlighting the method’s potential for improving visual quality in low-light conditions. Our code is available at <a href="https://github.com/ethanolallergy/CPMLED">https://github.com/ethanolallergy/CPMLED</a>.</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

Enhancing low-light images with simultaneous deblurring: a global–local interaction approach

  • 期刊论文

  • 发表日期：

    2026年3月29日

查看全文
  • Qiang Guo,
  • Zechen Wei,
  • Hongguang Pan,
  • Chaoxiu Yao,
  • Ze Jiang,
  • Libin Zhang

摘要

Low-light photography often results in degraded images with poor visibility and motion blur. Conventional methods typically address enhancement and deblurring sequentially, which can accumulate errors and yield suboptimal restoration results. We propose CPM-LED, an end-to-end network featuring a Curve Complementary Attention module for balanced enhancement, a Parallel Hybrid Attention module for deep feature representation, and a Multi-Scale Feature Enhancement module for effective deblurring. Experimental results demonstrate competitive performance with a PSNR of 27.481, SSIM of 0.892, and LPIPS of 0.202, highlighting the method’s potential for improving visual quality in low-light conditions. Our code is available at https://github.com/ethanolallergy/CPMLED.