Scholarlink

This study addresses Alzheimer’s disease detection through MRI scans, employing the renowned ResNet50, CNN with its distinctive 50 (48 convolutional layers, one MaxPool layer, and one average pool layer). ResNet50 is navigated using the Gaussian-based Bayesian Parameter Optimization (GBPO) technique for hyper-parameter optimization. This innovative approach yields optimal configurations for the network. To validate our model, we plan to compare its performance with the default parameter that optimizes are set to. When applied with the best parameters obtained from GBPO, these outcomes underscore the effectiveness and robustness of our optimized ResNet50 in Alzheimer’s disease detection. This paper serves as a concise illustration of the application of the proposed technique within a broader context, showcasing its adaptability to other deep convolutional neural networks (CNNs) such as Inception V3 and VGG19, DenseNet-169, LeNet-5, and AlexNet.

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 Gaussian–Bayesian Parameter Optimization-Based Deep Convolution Neural Network

  • 章节

  • 发表日期：

    2026年

查看全文
  • Ansh Varshney,
  • Ishita Luharuka,
  • Sobhan Sarkar,
  • Indranil Bose

摘要

This study addresses Alzheimer’s disease detection through MRI scans, employing the renowned ResNet50, CNN with its distinctive 50 (48 convolutional layers, one MaxPool layer, and one average pool layer). ResNet50 is navigated using the Gaussian-based Bayesian Parameter Optimization (GBPO) technique for hyper-parameter optimization. This innovative approach yields optimal configurations for the network. To validate our model, we plan to compare its performance with the default parameter that optimizes are set to. When applied with the best parameters obtained from GBPO, these outcomes underscore the effectiveness and robustness of our optimized ResNet50 in Alzheimer’s disease detection. This paper serves as a concise illustration of the application of the proposed technique within a broader context, showcasing its adaptability to other deep convolutional neural networks (CNNs) such as Inception V3 and VGG19, DenseNet-169, LeNet-5, and AlexNet.