Scholarlink

<p>Medical image fusion through multiple modalities functions as an essential tool for improved diagnosis because it harmonizes varied imaging results together. The research introduces an enhanced image fusion solution which unites Nonsubsampled Shearlet Transform (NSST) with Coupled Neural P (CNP) Systems to maximize the fusion results between MRI and PET brain images. The proposed system performs image decomposition starting from low and high frequency regions before it applies CNP systems for low-frequency fusion and WF-SML for high-frequency fusion. The proposed method proves superior in merging MRI and PET brain image pairs through experimental testing of 48 pairs when evaluated against seven mainstream fusion methods including CNN-based and neuro-fuzzy models. The proposed methodology grants improvements of 19.2% on Structural Similarity Index (SSIM) as well as 17.8% additional entropy while delivering improved standard deviation values to provide the best possible contrast and texture preservation and information maintenance. Medical image fusion techniques using our method generate detailed observations about Alzheimer’s disease along with brain tumors and neurodegenerative conditions which help medical professionals make better choices. Enhanced medical capabilities in imaging result from this modern fusion approach which improves both perceptible quality and clinical interpretation of combined images.</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

Multimodal Medical Image Fusion in NSST Domain in Coupled Neural Systems

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年5月12日

查看全文
  • Vella Satyanarayana,
  • P. Mohanaiah

摘要

Medical image fusion through multiple modalities functions as an essential tool for improved diagnosis because it harmonizes varied imaging results together. The research introduces an enhanced image fusion solution which unites Nonsubsampled Shearlet Transform (NSST) with Coupled Neural P (CNP) Systems to maximize the fusion results between MRI and PET brain images. The proposed system performs image decomposition starting from low and high frequency regions before it applies CNP systems for low-frequency fusion and WF-SML for high-frequency fusion. The proposed method proves superior in merging MRI and PET brain image pairs through experimental testing of 48 pairs when evaluated against seven mainstream fusion methods including CNN-based and neuro-fuzzy models. The proposed methodology grants improvements of 19.2% on Structural Similarity Index (SSIM) as well as 17.8% additional entropy while delivering improved standard deviation values to provide the best possible contrast and texture preservation and information maintenance. Medical image fusion techniques using our method generate detailed observations about Alzheimer’s disease along with brain tumors and neurodegenerative conditions which help medical professionals make better choices. Enhanced medical capabilities in imaging result from this modern fusion approach which improves both perceptible quality and clinical interpretation of combined images.