Scholarlink

Alzheimer’s disease (AD) is a prevalent type of dementia affecting memory and lacks a cure. Although the symptoms can be delayed with medications, early diagnosis is critical. Physicians use Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) scans to identify brain irregularities. Researchers employ machine learning to enhance early AD detection using these modalities. This study assesses two major machine learning techniques, Support Vector Machine (SVM) and Convolutional Neural Network (CNN), based on 45 research articles published from 2018 to 2024. The research highlights the significant impact of feature extraction and validation methods on the model’s performance for effective AD diagnosis. It highlights the variable accuracy of CNN and SVM models, influenced by the data quality and the specific feature extraction and validation techniques employed. The combination of different modalities notably improves SVM classifier performance, achieving an accuracy of 98%. Notably, using the Region of Interest (ROI) feature extraction method results in significantly enhanced accuracy of 100%, compared to other methods when employing CNN.

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

Optimizing CNN and SVM for Alzheimer’s Diagnosis: Insights from Feature Extraction Methods

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Akhilesh Deep Arya,
  • Sourabh Singh Verma,
  • Prasun Chakrabarti,
  • Rimpy Bishnoi

摘要

Alzheimer’s disease (AD) is a prevalent type of dementia affecting memory and lacks a cure. Although the symptoms can be delayed with medications, early diagnosis is critical. Physicians use Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) scans to identify brain irregularities. Researchers employ machine learning to enhance early AD detection using these modalities. This study assesses two major machine learning techniques, Support Vector Machine (SVM) and Convolutional Neural Network (CNN), based on 45 research articles published from 2018 to 2024. The research highlights the significant impact of feature extraction and validation methods on the model’s performance for effective AD diagnosis. It highlights the variable accuracy of CNN and SVM models, influenced by the data quality and the specific feature extraction and validation techniques employed. The combination of different modalities notably improves SVM classifier performance, achieving an accuracy of 98%. Notably, using the Region of Interest (ROI) feature extraction method results in significantly enhanced accuracy of 100%, compared to other methods when employing CNN.