BRAIN TUMOR DETECTION USING MRI IMAGES USING CONVOLUTIONAL NEURAL NETWORKS

Abstract: Serious neurological disorders marked by aberrant cell development inside the brain resulting from diverse genetic, environmental, or medicinal causes are brain tumors. Ignorance of these malignancies or treatment can greatly raise morbidity and death rates. Historically, bio-imaging technologies—among which Magnetic Resonance Imaging (MRI) is a generally acknowledged and non-invasive diagnostic tool—have been used in clinical assessment and diagnosis of brain tumors. Understanding the urgent need for early and accurate diagnosis, this proposed work focuses on building a Deep Learning Architecture (DLA) for automated brain tumor identification from two-dimensional MRI slices. The approach extract deep characteristics from MRI images using a pre-trained convolutional neural network—more especially, VGG19. A SoftMax classifier then sorts these derived features to either classify tumor existence or absence. Using transfer learning and deep feature extraction, the system seeks to accelerate diagnosis, lower human error, and enhance medical decision-making by The work also includes a structured MRI dataset specifically meant to facilitate model training, testing, and evaluation for strong performance results. Key Words: Brain Tumor, MRI (Magnetic Resonance Imaging), Deep Learning Architecture (DLA), VGG19, SoftMax Classifier, Feature Extraction