Seeing the Unseen - Deep Learning Approach for the Detection and Analysis of Brain Tumor

Detecting brain tumors quickly and accurately is essential for successful treatment and better patient prognosis. This project presents an integrated system capable of two tasks: segmenting tumor regions in MRI scans and identifying tumor type. The segmentation module uses a U-Net deep learning model trained to highlight affected areas in MRI images. Meanwhile, the classification part applies a VGG16-based CNN to categorize tumors into glioma, meningioma, or pituitary types. Through an intuitive web-based interface, users can choose to run segmentation or classification. After uploading MRI images, they receive either a visual mask showing tumor regions or the predicted tumor type. By combining deep learning with usability, the tool is designed to support healthcare professionals in faster and more reliable brain tumor analysis.