Scholarlink

<p>Brain tumors represent a significant health concern worldwide, and early diagnosis through medical imaging techniques such as MRI plays a critical role in effective treatment planning. Manual tumor identification using magnetic resonance imaging (MRIs) is time-consuming and falls short of reliably detecting, localizing, and categorizing tumor types. The advances in computer vision and machine learning algorithms can quickly detect and categorize tumors from MRI images without the help of medical specialists. This research work proposes an innovative brain image segmentation technique using Kinetic Gas Molecular Optimized Weighted Fuzzy C-Means (KGMO-WFCM) with trained UNET. Here, the brain regions like Grey Matter (GM), Cerebrospinal Fluid (CSF), and White Matter (WM) are segmented using the proposed technique. The modified KGMO-WFCM technique with trained UNET helps to separate GM, WM and CSF region which reduces the computational complexity. Finally, the tumor region is segmented with high accuracy. Various performance metrics like accuracy, specificity, sensitivity, Jaccard Coefficient (JAC), Hausdorff Distance (HD) and Dice Similarity Coefficient (DC) are calculated to validate the performance of the proposed work. Proposed method produce 98.34% of accuracy, 96.46% of sensitivity, 99.87% of specificity, 0.063 of JAC, 12.03 of HD, 99.64% of DC and 61.24% of Running time Reduction Rate and outperforms better than the existing methods.</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

Robust MRI Image Segmentation: Leveraging the Power of KGM Optimized W-FCM with Deep Learning

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Shenbagarajan Anantharajan,
  • Shenbagalakshmi Gunasekaran,
  • Sivaganesh Anantharajan,
  • Thavasi Subramanian

摘要

Brain tumors represent a significant health concern worldwide, and early diagnosis through medical imaging techniques such as MRI plays a critical role in effective treatment planning. Manual tumor identification using magnetic resonance imaging (MRIs) is time-consuming and falls short of reliably detecting, localizing, and categorizing tumor types. The advances in computer vision and machine learning algorithms can quickly detect and categorize tumors from MRI images without the help of medical specialists. This research work proposes an innovative brain image segmentation technique using Kinetic Gas Molecular Optimized Weighted Fuzzy C-Means (KGMO-WFCM) with trained UNET. Here, the brain regions like Grey Matter (GM), Cerebrospinal Fluid (CSF), and White Matter (WM) are segmented using the proposed technique. The modified KGMO-WFCM technique with trained UNET helps to separate GM, WM and CSF region which reduces the computational complexity. Finally, the tumor region is segmented with high accuracy. Various performance metrics like accuracy, specificity, sensitivity, Jaccard Coefficient (JAC), Hausdorff Distance (HD) and Dice Similarity Coefficient (DC) are calculated to validate the performance of the proposed work. Proposed method produce 98.34% of accuracy, 96.46% of sensitivity, 99.87% of specificity, 0.063 of JAC, 12.03 of HD, 99.64% of DC and 61.24% of Running time Reduction Rate and outperforms better than the existing methods.