A Deep Convolutional Neural Network Based Architecture for Accurate Detection of Brain Diseases Using Medical Imaging

Abstract

Brain tumors are among the most life-threatening neurological disorders, and their early, accurate diagnosis through Magnetic Resonance Imaging (MRI) is critical for effective treatment planning. Manual interpretation of MRI scans is time-consuming, subjective, and prone to inter-observer variability, particularly in regions with a shortage of trained neuro-radiologists. This paper presents a deep Convolutional Neural Network (CNN) based architecture, referred to as NeuroScan-Net, for the automated detection and classification of brain diseases from MRI images into four categories: Glioma Tumor, Meningioma Tumor, Pituitary Tumor, and No Tumor (Normal). The proposed system combines an image preprocessing and augmentation pipeline with a custom CNN comprising four convolutional blocks with batch normalization and max-pooling, followed by a fully connected classification head with dropout regularization and a softmax output layer. The model is trained and evaluated on a combined, publicly available brain MRI dataset containing 7,023 T1-weighted contrast-enhanced images spanning the four target classes. Experimental evaluation indicates that the proposed lightweight architecture achieves classification accuracy competitive with much larger pre-trained transfer-learning networks such as VGG16, ResNet50 and InceptionV3, while requiring substantially fewer trainable parameters and lower inference latency. An interactive web-based diagnostic interface was also developed to demonstrate real-time inference, confidence scoring, and clinical decision-support information for each predicted class. The results suggest that the proposed CNN architecture is a viable, computationally efficient tool for assisting radiologists in the preliminary screening of brain tumors from MRI scans.