Scholarlink

<p>Alzheimer’s disease (AD) is an advancing neurodegenerative condition marked by cognitive decline and functional impairment. Its early symptoms often develop gradually and are difficult to detect, making timely diagnosis a significant challenge. Accurate early-stage classification is essential for initiating interventions that may delay disease progression and enhance patients’ quality of life. Traditional diagnostic techniques are resource-intensive and often lack scalability or interpretability. This research aims to develop a deep learning-based classification model that can identify and distinguish between the stages of Alzheimer’s disease using MRI scans, with a focus on achieving high accuracy and clinical interpretability. We introduce ADNet, a deep learning architecture that utilises optimised Convolutional Neural Networks (CNNs) combined with explainable AI (XAI) techniques. The model is trained on a balanced and diverse dataset of 4000 MRI scans curated from the publicly available Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. ADNet achieved a classification accuracy of 99.40% and an Area Under the Curve (AUC) of 99.85% on the test dataset. Grad-CAM was applied for model interpretability, visually highlighting the brain regions contributing to predictions. The proposed ADNet model is a promising diagnostic tool for Alzheimer’s disease, offering both high accuracy and interpretability. Its novelty lies in the integration of a finely tuned CNN architecture with visual interpretability through Grad-CAM, enabling not only precise stage-wise classification but also transparent insights into model decision-making. This combination of performance and explainability addresses key challenges in clinical adoption and supports more informed and timely diagnostic decisions in real-world settings.</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

Transforming Alzheimer’s diagnosis: ADNet deep learning with explainable AI framework

  • 期刊论文

  • 发表日期：

    2025年8月12日

查看全文
  • Anant Singhal,
  • A. Charan Kumari,
  • K. Srinivas

摘要

Alzheimer’s disease (AD) is an advancing neurodegenerative condition marked by cognitive decline and functional impairment. Its early symptoms often develop gradually and are difficult to detect, making timely diagnosis a significant challenge. Accurate early-stage classification is essential for initiating interventions that may delay disease progression and enhance patients’ quality of life. Traditional diagnostic techniques are resource-intensive and often lack scalability or interpretability. This research aims to develop a deep learning-based classification model that can identify and distinguish between the stages of Alzheimer’s disease using MRI scans, with a focus on achieving high accuracy and clinical interpretability. We introduce ADNet, a deep learning architecture that utilises optimised Convolutional Neural Networks (CNNs) combined with explainable AI (XAI) techniques. The model is trained on a balanced and diverse dataset of 4000 MRI scans curated from the publicly available Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. ADNet achieved a classification accuracy of 99.40% and an Area Under the Curve (AUC) of 99.85% on the test dataset. Grad-CAM was applied for model interpretability, visually highlighting the brain regions contributing to predictions. The proposed ADNet model is a promising diagnostic tool for Alzheimer’s disease, offering both high accuracy and interpretability. Its novelty lies in the integration of a finely tuned CNN architecture with visual interpretability through Grad-CAM, enabling not only precise stage-wise classification but also transparent insights into model decision-making. This combination of performance and explainability addresses key challenges in clinical adoption and supports more informed and timely diagnostic decisions in real-world settings.