Scholarlink

Alzheimer's Disease (AD) is a type of dementia that progresses with time. This disease impacts cognitive functions of elderly patients affecting their daily life. This paper provides a cutting-edge approach to early AD detection using advanced deep learning system, specifically targeting the challenge of timely and accurate diagnosis. This is done by feeding neurodiagnostic imaging data, especially Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) scans, through the proposed hybrid 3D CNN-LSTM model. The methodology encompasses comprehensive data preprocessing, including image normalization, augmentation, and skull stripping, to enhance model performance. The research achieves a hybrid 3D CNN-LSTM model with 94% accuracy in classifying Alzheimer's disease stages, successfully processing and analyzing neuroimaging data from the ADNI dataset. The proposed model proves to be effective in handling spatial data such as shape, size, position etc. as well as temporal data such as trends, patterns and sequences. The study enables precise early detection across four cognitive stages: Alzheimer's Disease (AD), Cognitive Normal (CN), Early Mild Cognitive Impairment (EMCI), and Mild Cognitive Impairment (MCI). A user-friendly web interface further supports scan upload and diagnostic classification, contributing a promising computational methodology for identifying early onset of Alzheimer’s Disease and potentially revolutionizing diagnostic approaches in clinical practice.

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

Early Prediction of Alzheimer’s Disease Using Deep Learning Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • S. Deeparani,
  • G. Archana,
  • R. Deepika,
  • D. Dhanrithii

摘要

Alzheimer's Disease (AD) is a type of dementia that progresses with time. This disease impacts cognitive functions of elderly patients affecting their daily life. This paper provides a cutting-edge approach to early AD detection using advanced deep learning system, specifically targeting the challenge of timely and accurate diagnosis. This is done by feeding neurodiagnostic imaging data, especially Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) scans, through the proposed hybrid 3D CNN-LSTM model. The methodology encompasses comprehensive data preprocessing, including image normalization, augmentation, and skull stripping, to enhance model performance. The research achieves a hybrid 3D CNN-LSTM model with 94% accuracy in classifying Alzheimer's disease stages, successfully processing and analyzing neuroimaging data from the ADNI dataset. The proposed model proves to be effective in handling spatial data such as shape, size, position etc. as well as temporal data such as trends, patterns and sequences. The study enables precise early detection across four cognitive stages: Alzheimer's Disease (AD), Cognitive Normal (CN), Early Mild Cognitive Impairment (EMCI), and Mild Cognitive Impairment (MCI). A user-friendly web interface further supports scan upload and diagnostic classification, contributing a promising computational methodology for identifying early onset of Alzheimer’s Disease and potentially revolutionizing diagnostic approaches in clinical practice.