Scholarlink

In the fight against Alzheimer’s disease and its associated cognitive decline, our team has crafted an automated system for the analysis of brain MRI scans, emphasizing accurate dementia staging. Utilizing transfer learning techniques with the ResNet model on data from the KAGGLE dataset, our methodology showcases an exceptional 95.12% accuracy in multiclass classifications. This level of precision, augmented by performance metrics such as 96.13% specificity, 82.14% precision, 96.12% recall, and an 83.32% F1 score, affirms the system’s reliability and precision in diagnosing various stages of dementia. Additionally, the system’s ROC curve performance, especially an AUC of 0.97964 for “Moderate Dementia,” confirms its early and accurate disease detection and staging capability. By setting a new benchmark in the early diagnosis and precise staging of Alzheimer’s disease, our approach holds promise for significantly impacting patient care and advancing research in neurodegenerative disease management.

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

Advancing Alzheimer’s Diagnosis: Integrating Transfer Learning with Support Vector Machine for Enhanced Disease Classification

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mrutyunjaya S. Hiremath,
  • Rajashekhar C. Biradar,
  • N. Praveen,
  • Jayaprada S. Hiremath,
  • Sujith Kumar,
  • Shantala S. Hiremath

摘要

In the fight against Alzheimer’s disease and its associated cognitive decline, our team has crafted an automated system for the analysis of brain MRI scans, emphasizing accurate dementia staging. Utilizing transfer learning techniques with the ResNet model on data from the KAGGLE dataset, our methodology showcases an exceptional 95.12% accuracy in multiclass classifications. This level of precision, augmented by performance metrics such as 96.13% specificity, 82.14% precision, 96.12% recall, and an 83.32% F1 score, affirms the system’s reliability and precision in diagnosing various stages of dementia. Additionally, the system’s ROC curve performance, especially an AUC of 0.97964 for “Moderate Dementia,” confirms its early and accurate disease detection and staging capability. By setting a new benchmark in the early diagnosis and precise staging of Alzheimer’s disease, our approach holds promise for significantly impacting patient care and advancing research in neurodegenerative disease management.