Scholarlink

<p>Alzheimer’s disease (AD) is an irreversible neurological disorder that causes a steady deterioration in cognitive function. Since AD develops years before symptoms manifest and biomarkers, which are mostly detectable by a variety of neuroimaging modalities, vary subtlety, early detection is challenging. To overcome these challenges, this study proposes a unique Hybrid Structural Graph Discrete Hopfield Neural Network with Doll Maker Optimization (HSGDHNN-DMO) to detect AD in MRI scans. First, MRI pictures are gathered from the OASIS and ADNI datasets. Subsequently, An Adaptive Mesh Denoising Method with GCN-Based Guided Normal Filtering (GNF-Net) is an enhancement of the adaptive mesh denoising technique that removes noise, normalizes intensity, and protects important anatomical structures. Following that, segmentation is done using Multi-View Fuzzy Clustering based on Anchor Graph (MVFCAG), which creates balanced, unique clusters by combining multi-view data and trace norm regularization. Then, Hybrid Structural Graph Discrete Hopfield Neural Network (HSGDHNN) combines memory-preserving dynamics with spatial-semantic attention to capture intricate pathological patterns typical of Alzheimer’s disease. Finally, Doll Maker Optimization (DMO) algorithm is applied for hyperparameter tuning, leveraging a novel exploration–exploitation strategy inspired by individual pattern design. The proposed approach is quite accurate for medical diagnosis since it significantly improves AD detection while lowering false positives. It achieves remarkable accuracy (99.82% on ADNI and 99.84% on OASIS), precision (99.68% on ADNI and 99.68% on OASIS), and F1-score (99.69% on ADNI and 99.73% on OASIS).</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

Alzheimer’s Disease Detection Using Hybrid Structural Graph Discrete Hopfield Neural Network with Doll Maker Optimization Algorithm

  • 期刊论文

  • 发表日期：

    2025年8月5日

查看全文
  • Tripti Sharma,
  • Nidhya Rangarajan,
  • Sreevardhan Cheerla,
  • Krishna Prakash Arunachalam

摘要

Alzheimer’s disease (AD) is an irreversible neurological disorder that causes a steady deterioration in cognitive function. Since AD develops years before symptoms manifest and biomarkers, which are mostly detectable by a variety of neuroimaging modalities, vary subtlety, early detection is challenging. To overcome these challenges, this study proposes a unique Hybrid Structural Graph Discrete Hopfield Neural Network with Doll Maker Optimization (HSGDHNN-DMO) to detect AD in MRI scans. First, MRI pictures are gathered from the OASIS and ADNI datasets. Subsequently, An Adaptive Mesh Denoising Method with GCN-Based Guided Normal Filtering (GNF-Net) is an enhancement of the adaptive mesh denoising technique that removes noise, normalizes intensity, and protects important anatomical structures. Following that, segmentation is done using Multi-View Fuzzy Clustering based on Anchor Graph (MVFCAG), which creates balanced, unique clusters by combining multi-view data and trace norm regularization. Then, Hybrid Structural Graph Discrete Hopfield Neural Network (HSGDHNN) combines memory-preserving dynamics with spatial-semantic attention to capture intricate pathological patterns typical of Alzheimer’s disease. Finally, Doll Maker Optimization (DMO) algorithm is applied for hyperparameter tuning, leveraging a novel exploration–exploitation strategy inspired by individual pattern design. The proposed approach is quite accurate for medical diagnosis since it significantly improves AD detection while lowering false positives. It achieves remarkable accuracy (99.82% on ADNI and 99.84% on OASIS), precision (99.68% on ADNI and 99.68% on OASIS), and F1-score (99.69% on ADNI and 99.73% on OASIS).