Scholarlink

<p>Precise segmentation of brain tumors in MRI scans is essential for accurate diagnosis, treatment planning, and monitoring of disease progression. To overcome challenges related to both data security and segmentation accuracy, this study introduces a novel framework that integrates blockchain technology with machine learning. We propose an enhanced architecture that combines a modified Edge U-Net model augmented with Edge Guidance Blocks (EGBs) and LeakyReLU activation functions for improved feature extraction and boundary localization with a blockchain-based data management system. The framework operates on multimodal MRI data securely stored and verified through a blockchain platform, ensuring data integrity, provenance, and tamper resistance. To further optimize performance, a Lightweight Proof-of-Work (LPoW) consensus algorithm is introduced, significantly reducing block processing time and communication overhead compared to traditional consensus mechanisms. Experimental evaluations confirm that the proposed approach outperforms existing state-of-the-art models, achieving 99.71% accuracy, 92.83% Dice, and 94.50% IoU on the BraTS dataset. In addition, the LPoW consensus significantly lowered block processing time, achieving 2.3× and 3.2× faster performance compared to PoW and PBFT.</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

Blockchain-enhanced brain tumor prediction: a novel approach leveraging machine learning

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • Vijayant Pawar,
  • Shelly Sachdeva

摘要

Precise segmentation of brain tumors in MRI scans is essential for accurate diagnosis, treatment planning, and monitoring of disease progression. To overcome challenges related to both data security and segmentation accuracy, this study introduces a novel framework that integrates blockchain technology with machine learning. We propose an enhanced architecture that combines a modified Edge U-Net model augmented with Edge Guidance Blocks (EGBs) and LeakyReLU activation functions for improved feature extraction and boundary localization with a blockchain-based data management system. The framework operates on multimodal MRI data securely stored and verified through a blockchain platform, ensuring data integrity, provenance, and tamper resistance. To further optimize performance, a Lightweight Proof-of-Work (LPoW) consensus algorithm is introduced, significantly reducing block processing time and communication overhead compared to traditional consensus mechanisms. Experimental evaluations confirm that the proposed approach outperforms existing state-of-the-art models, achieving 99.71% accuracy, 92.83% Dice, and 94.50% IoU on the BraTS dataset. In addition, the LPoW consensus significantly lowered block processing time, achieving 2.3× and 3.2× faster performance compared to PoW and PBFT.