Speed of Diagnosis for Brain Diseases Using MRI and Convolutional Neural Networks

Accurately diagnosing brain diseases is crucial for effective treatment and improved patient outcomes. Magnetic Resonance Imaging is a regularly used technology in the investigation of brain illnesses including Alzheimer's disease, brain tumors, and multiple sclerosis. This study proposes a Convolutional Neural Network-based automated brain illness classification method utilizing MRI images. The proposed method leverages a dataset of MRI images of four brain diseases, namely Alzheimer’s disease, tumors of brain, multiple sclerosis, and healthy brains. We trained and compared different CNN architectures, including VGG16 and fine-tuned ResNet. Our CNN model achieved remarkable accuracy on both the training and testing sets. Specifically, we achieved an impressive training accuracy of 99.01% and a testing accuracy of 95%, outperforming VGG16 and fine-tuned ResNet. We derived many assessment measures, including accuracy, recall, and F1-score, to further evaluate the effectiveness of our model. Our results demonstrate the potential of CNN-based approaches in accurately and automatically classifying brain diseases using MRI images. Our proposed approach has the potential to be a valuable tool for healthcare professionals, improving patient outcomes and quality of life. The developed model is capable of classifying Alzheimer’s disease, brain tumors, multiple sclerosis, and their respective stages. Automated classification of brain diseases using CNNs could enable early detection and precise diagnosis of these diseases, leading to improved treatment and patient care.