Brain Disease Detection and Health Insights Using Deep Learning

This paper presents a novel approach for brain disease detection and health insights using deep learning, specifically leveraging ResNet-50, a pre-trained convolutional neural network model. The system classifies brain conditions into three categories: Alzheimer’s, Mild Cognitive Impairment, and Healthy/Normal, based on MRI brain images. The classified output is then utilized to estimate a range of brain age and provide personalized recommendations and precautions. The methodology involves preprocessing MRI images, resizing them to 224x224 dimensions, and ensuring RGB format compatibility. Using a curated dataset, the images are augmented and split into training and validation sets. The ResNet-50 model is fine-tuned by modifying the fully connected layer to accommodate three output classes, followed by training with stochastic gradient descent momentum (SGDM) optimization. The system’s performance is evaluated using accuracy, precision, recall, and F1-score metrics. Additionally, a user-friendly graphical user interface (GUI) facilitates MRI image upload, classification output display, age estimation, and corresponding health recommendations. The results demonstrate the model's efficacy in accurately classifying brain diseases, with potential applications in early diagnosis and cognitive health monitoring. This approach not only aids healthcare professionals in decision-making but also empowers patients with proactive health management strategies, thus contributing to improved healthcare outcomes.

Keywords: Convolutional neural network, Deep learning, ResNet-50, image Processing techniques and Classification.