AI-Driven Brain Tumor Detection based on Convolutional Neural Networks

Brain Tumor can lead to numerous health issues such as headache, vomiting and drowsiness, nausea (i.e. persistently feeling sick), seizures. Also can lead to mental or behavioural changes, memory problems or changes in personality. In certain conditions, can lead to progressive weakness which can cause paralysis on one side of the body vision or speech problems, etc. If we detect Brain Tumor in early stage then it could save life and treatment would be easier. For the detection of brain Tumor, built a CNN model. Model takes 39.5 MB of storage. Model can classify between 4 classes namely – No tumor, Glioma, Meningioma and Pituitary. Images that are used to train the model are of MRI (i.e. Magnetic Resonance Imaging) scans as they are the commonly used for detecting brain tumor. The model has training accuracy of 99.20%, validation accuracy of 97.68% and testing accuracy of 97.10%.