Brain Tumor and Alzheimer Detection using CNN

Neurological disorders including brain tumors and Alzheimer's disease impose a substantial global healthcare burden, necessitating reliable and timely diagnostic tools. This paper presents a deep learning framework grounded in Convolutional Neural Networks (CNNs) for simultaneous multi-class classification of brain MRI images across two distinct domains: four categories of brain tumors (Glioma, Meningioma, Pituitary Tumor, and No Tumor) and four progression stages of Alzheimer's disease (Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented). Input images underwent standardized preprocessing including spatial resizing to 224×224 pixels, pixel-intensity normalization, and augmentationbased regularization prior to training. Both CNN models were optimized using the Adam optimizer with categorical cross-entropy as the loss criterion. Experimental evaluation yielded test accuracies of 95.6% and 95.3% for tumor and Alzheimer's classification respectively, with macro-averaged F1-scores exceeding 94.7% across all classes. Confusion matrix analysis confirmed strong inter-class discriminability with negligible misclassification. Both trained models were integrated within a Django-based web application enabling real-time clinical inference from raw MRI uploads. The outcomes affirm the viability of data-driven CNN pipelines as decision-support aids for neurological diagnostics.