Scholarlink

<p>Compressed sensing (CS) presents an innovative method for acquiring high-resolution images in Computed Tomography (CT) with fewer measurements. This approach not only accelerates scan times but also has the potential to lower radiation exposure for patients. However, the quality of image reconstruction in CS is highly dependent on the coherence properties of the sensing matrix. This paper aims to enhance image reconstruction quality in CS medical imaging by introducing a novel approach that improves the Joint Optimization algorithm for Average mutual coherence and Cumulative coherence (JOAC) via preconditioning. This approach simplifies threshold selection and reduces computation time by incorporating an updated shrinkage function. Experimental results demonstrate that our proposed optimization algorithm significantly improves the reconstruction quality of medical images. In noiseless scenarios, we observed improvements of 10 dB in Peak Signal-to-Noise Ratio (PSNR), and in noisy scenarios, we achieved a 4 dB improvement in PSNR, alongside a reduction in reconstruction time, further enhancing the efficiency of CS-based CT imaging.</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

Enhanced image reconstruction in compressed sensing medical imaging via preconditioning

  • 期刊论文

  • 发表日期：

    2025年8月8日

查看全文
  • Evelin Nissy Thomas,
  • Prasad Theeda,
  • Praveen Thomas

摘要

Compressed sensing (CS) presents an innovative method for acquiring high-resolution images in Computed Tomography (CT) with fewer measurements. This approach not only accelerates scan times but also has the potential to lower radiation exposure for patients. However, the quality of image reconstruction in CS is highly dependent on the coherence properties of the sensing matrix. This paper aims to enhance image reconstruction quality in CS medical imaging by introducing a novel approach that improves the Joint Optimization algorithm for Average mutual coherence and Cumulative coherence (JOAC) via preconditioning. This approach simplifies threshold selection and reduces computation time by incorporating an updated shrinkage function. Experimental results demonstrate that our proposed optimization algorithm significantly improves the reconstruction quality of medical images. In noiseless scenarios, we observed improvements of 10 dB in Peak Signal-to-Noise Ratio (PSNR), and in noisy scenarios, we achieved a 4 dB improvement in PSNR, alongside a reduction in reconstruction time, further enhancing the efficiency of CS-based CT imaging.