Early Detection
Saves Lives.
AI Makes It Possible.

NeuraScan uses deep learning (MobileNetV2) to analyze brain MRI scans and classify tumor types with clinical-grade accuracy — helping researchers and medical professionals make faster, better-informed decisions.

94%
Accuracy
4
Tumor Classes
MobileNetV2
Architecture
Transfer Learning
Learning Method
⚠ GLIOMA DETECTED
✓ ANALYSIS READY
Confidence Score
94.2%
About the Project

AI-Assisted Brain
Tumor Analysis

NeuraScan is a Final Year Project developed at KFUEIT, combining transfer learning with MobileNetV2 to build a clinically-relevant tool for brain tumor detection from MRI images.

Transfer Learning ArchitectureLeverages pre-trained MobileNetV2 weights on ImageNet, fine-tuned on a curated brain tumor MRI dataset.
4-Class ClassificationDistinguishes between Glioma, Meningioma, Pituitary tumor, and No Tumor with a single forward pass.
Real-time & Upload ReadyOptimized pipeline for both static MRI upload and real-time processing.
Flask-Powered BackendLightweight Python web server handles inference and result rendering efficiently.
Classification

4 Categories Detected

Our model identifies and classifies the most clinically significant brain tumor types

Glioma
High-Grade Malignant

Originates from glial cells. Most common and aggressive brain tumor.

Meningioma
Mostly Benign

Arises from the meninges surrounding the brain. Often slow-growing.

Pituitary
Hormonal Impact

Affects the pituitary gland and hormonal regulation.

No Tumor
Clear Scan

No evidence of tumor. High specificity minimizes false positives.

Technology Stack

Built with Modern AI Tools

Production-grade technologies powering every layer of NeuraScan

MobileNetV2
Lightweight CNN architecture fine-tuned for brain tumor classification.
TensorFlow / Keras
Model training, validation, and serialization.
Flask Backend
Serving the ML model via REST API.
OpenCV
Real-time webcam capture and preprocessing.
NumPy / Pillow
Image array manipulation and augmentation utilities.
HuggingFace Spaces
Cloud deployment with zero-config GPU inference.
Get Started

Ready to Detect Tumors
with AI Precision?

Create a free account and start uploading MRI scans for instant AI-powered classification. Built for researchers, medical students, and healthcare professionals.

Create Free Account Already have an account?