🧠 NeuroScan AI Upload MRI
AI-Powered Medical Imaging

NeuroScan AI

AI-Powered Brain Tumor Detection System

Upload MRI scans and instantly detect brain tumors using a custom Convolutional Neural Network.

✓ Glioma ✓ Meningioma ✓ Pituitary Tumor ✓ No Tumor
86.25%
Test Accuracy
7,200 MRI Images Trained

Core Features

Enterprise-grade AI medical imaging tools available directly in your browser

🧠

Deep Learning CNN

Custom Convolutional Neural Network trained on 7,200 MRI images

🏥

Medical AI

AI-assisted brain tumor classification using clinical-grade MRI scans

Real-Time Prediction

Instant classification with detailed per-class confidence scores

🎯

High Accuracy

86.25% test accuracy with 94.09% training accuracy across 4 classes

How It Works

Three simple steps to get your MRI analysis result

📤
Step 1

Upload MRI Scan

Drag and drop or browse to select your MRI image (JPG, JPEG, PNG)

🤖
Step 2

AI Analysis

Our CNN model processes the image through trained convolutional layers

📊
Step 3

Get Results

Receive prediction label and per-class confidence scores instantly

Upload MRI Scan

Select or drag and drop your brain MRI image for instant AI analysis

🖼️

Drop MRI Image Here

or

Browse File

Supported: JPG · JPEG · PNG

Model Performance

Real metrics from training and evaluation on 7,200 MRI images

7,200
Dataset Size
94.09%
Train Accuracy
86.25%
Test Accuracy
4
Classes
Custom CNN
Architecture

Training Visualizations

📈

Accuracy & Loss Curves

Training and Validation Accuracy and Loss Curves
Train Validation
🔢

Confusion Matrix

Confusion Matrix Brain Tumor Classification
Rows = Actual Class  ·  Columns = Predicted Class

Tech Stack

Built with industry-standard ML and data science tools

🐍
Python
Language
🟦
TensorFlow
Deep Learning
⚙️
Keras
Neural Networks
👁️
OpenCV
Image Processing
🔢
NumPy
Numerical Computing
Vercel
Deployment
📉
Matplotlib
Visualization
📐
Scikit-Learn
ML Utilities

About This Project

A deep learning project for medical image classification

Project Overview

This project uses a custom Convolutional Neural Network (CNN) to classify brain MRI scans into four categories. Built to demonstrate practical applications of deep learning in the medical imaging domain.

The model was trained on 7,200 MRI images and achieves 86.25% test accuracy, making it a strong proof-of-concept for AI-assisted diagnostics.

Classified Conditions

Glioma
Tumor occurring in the brain and spinal cord, arising from glial cells
Meningioma
Tumor arising from the meninges, the membranes surrounding the brain
Pituitary Tumor
Abnormal growth in the pituitary gland at the base of the brain
No Tumor
Normal MRI scan with no detectable tumor present