Scholarlink

<p>The White Matter (WM) segmentation in magnetic resonance imaging (MRI) is crucial for the diagnosis and tracking of neurological illnesses since WM volume loss is a key biomarker for multiple sclerosis, Alzheimer’s disease, and other neurodegenerative diseases. Low contrast between WM and adjacent tissues, intrinsic image noise, and the possibility of losing structural detail during image enhancement make MRI-based WM segmentation difficult. These restrictions may impair early illness diagnosis and lower diagnostic reliability. In order to tackle these problems, this study presents a new segmentation framework that combines a Pseudo-Trapezoidal Membership-based Spatial Fuzzy Clustering Algorithm (PMSFCA) for precise tissue classification, Feature-preserving Contrast Enhancement Transform (FCET) for better visibility of WM boundaries, and Locally-Adaptive Steering Kernel Regression (LASKR) for simultaneous de-noising and sharpening. By combining noise reduction and enhancement, this integrated method lowers computing cost, produces accurate WM segmentation, and enhances picture contrast and sharpness without adding artifacts. The suggested approach provides a reliable, effective, and practically applicable tool for improving the diagnosis of neurological diseases. The remarkable capacity of the suggested algorithm LASKR to smooth noise while preserving fine structural features is demonstrated by its lowest MLV value of 0.0634 ± 0.0041. The fact that it had the lowest SD (0.0676 ± 0.0158) further demonstrated its stability under various picture situations. With a computational efficiency of 1.5816 ± 0.17 seconds, it was the quickest method, which makes it appropriate for real-time applications. PSNR (37.2 dB) and PMSE (0.014), two performance measurements, further confirm its capacity to generate high-quality pictures.</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

Joint Denoising and Sharpening of MR Imagery Using Locally-Adaptive Steering Kernel Regression

  • 期刊论文

  • 发表日期：

    2026年2月26日

查看全文
  • S. VinuRajkumar,
  • S. Sathish,
  • N. Malathy,
  • N. Indumathi

摘要

The White Matter (WM) segmentation in magnetic resonance imaging (MRI) is crucial for the diagnosis and tracking of neurological illnesses since WM volume loss is a key biomarker for multiple sclerosis, Alzheimer’s disease, and other neurodegenerative diseases. Low contrast between WM and adjacent tissues, intrinsic image noise, and the possibility of losing structural detail during image enhancement make MRI-based WM segmentation difficult. These restrictions may impair early illness diagnosis and lower diagnostic reliability. In order to tackle these problems, this study presents a new segmentation framework that combines a Pseudo-Trapezoidal Membership-based Spatial Fuzzy Clustering Algorithm (PMSFCA) for precise tissue classification, Feature-preserving Contrast Enhancement Transform (FCET) for better visibility of WM boundaries, and Locally-Adaptive Steering Kernel Regression (LASKR) for simultaneous de-noising and sharpening. By combining noise reduction and enhancement, this integrated method lowers computing cost, produces accurate WM segmentation, and enhances picture contrast and sharpness without adding artifacts. The suggested approach provides a reliable, effective, and practically applicable tool for improving the diagnosis of neurological diseases. The remarkable capacity of the suggested algorithm LASKR to smooth noise while preserving fine structural features is demonstrated by its lowest MLV value of 0.0634 ± 0.0041. The fact that it had the lowest SD (0.0676 ± 0.0158) further demonstrated its stability under various picture situations. With a computational efficiency of 1.5816 ± 0.17 seconds, it was the quickest method, which makes it appropriate for real-time applications. PSNR (37.2 dB) and PMSE (0.014), two performance measurements, further confirm its capacity to generate high-quality pictures.