Design of a CNN+GRU Transformer Model for Alzheimer’s Disease Prediction Using MRI Images

Alzheimer’s disease is one of the most prevalent neurodegenerative disorders, and early detection is essential for better management. This project proposes a hybrid deep learning model combining Convolutional Neural Networks (CNN), Swin Transformer, and Conformer Transformer, model combining with Convolutional Neural Networks (CNN)+GRU and Mobilenet for predicting Alzheimer’s disease using MRI images. The model utilizes CNN to extract meaningful features from the MRI scans, Swin Transformer for capturing fine-grained spatial information, and Conformer Transformer to model sequential dependencies in the data. The dataset used for training the model is obtained from Kaggle, containing MRI images labeled with different stages of Alzheimer’s disease. The proposed system aims to classify MRI images into the correct Alzheimer’s disease stages with high accuracy. This web-based system provides a user-friendly interface where medical professionals can upload MRI images and receive predictions on the likelihood of Alzheimer’s disease. By leveraging the power of deep learning models, the system aims to support healthcare providers in early diagnosis, improving patient outcomes.

Keywords: Alzheimer's disease, MRI images, deep learning, CNN, Swin Transformer, Conformer Transformer, GRU, Mobilenet prediction model, disease classification, machine learning, medical diagnosis