Multimodal Deep LearningBased onBrain Stroke Detection Using Imaging and LifestyleData

The project titled "Multimodal Deep Learning-Based Brain Stroke Detection Using Imaging and Lifestyle Data" aims to develop a web-based application for the early detection of brain strokes using a combination of medical imaging (ResNet50-based image classification) and lifestyle factors (including age, gender, hypertension, heart disease, smoking status, and glucose levels). The system utilizes both clinical and imaging data to make more accurate predictions, offering a robust method for identifying individuals at high risk of a stroke. The application features user authentication, file upload, model training, and image prediction capabilities. It allows users to upload both CSV datasets (containing lifestyle data) and medical images (such as X-rays or CT scans). Machine learning models, including Linear Discriminant Analysis (LDA), Gradient Boosting, Multi-Layer Perceptron (MLP), and Convolutional Neural Networks (CNN), are employed for prediction. The system offers a real-time prediction route for users to assess their stroke risk based on personal data, providing personalized health recommendations. Additionally, the image classification aspect of the project incorporates the ResNet model, trained to classify medical images into categories of stroke detection, offering insights based on the visual assessment of medical scans.

Keywords: Brain Stroke Detection, Multimodal Deep Learning, Imaging Data, Lifestyle Data, Machine Learning, ResNet, CSV Upload, Gradient Boosting, Linear Discriminant Analysis, Prediction System, Flask Web Application.