Get Started

This platform classifies brain MRI images into 4 categories using a ResNet50 deep learning model. Follow the steps below to run your first diagnosis.

1
Create an Account

Sign up with a username and password. Each user gets their own model space and inference history.

Sign Up
2
Train a Model

Go to Train Model, configure epochs/batch size/learning rate, and train ResNet50 on your MRI dataset.

Login required
3
Upload & Diagnose

Upload a brain MRI image (JPEG/PNG). The model returns the classification and confidence score in seconds.

Login required

Tumor Categories

The model classifies MRI images into these 4 categories. Below are placeholder examples -- replace them with real MRI snapshots from your dataset.

No Tumor example
No Tumor
Healthy
Glioma example
Glioma
Malignant
Meningioma example
Meningioma
Benign
Pituitary example
Pituitary
Benign
Dataset Structure

To train the model, provide a dataset with this folder structure:

Training/
  ├── notumor/
  │   ├── image001.jpg
  │   ├── image002.jpg
  │   └── ...
  ├── glioma/
  │   ├── image001.jpg
  │   └── ...
  ├── meningioma/
  │   ├── image001.jpg
  │   └── ...
  └── pituitary/
      ├── image001.jpg
      └── ...

Set the DATASET_PATH environment variable to point to your Training/ folder. Images should be brain MRI scans in JPEG or PNG format.

Technical Details
Model Architecture
  • Base: ResNet50 (pre-trained on ImageNet)
  • Input size: 224 x 224 px RGB
  • Output: 4 classes with softmax probabilities
  • Transfer learning with fine-tuning
Default Training Parameters
  • Epochs: 10
  • Batch size: 16
  • Learning rate: 0.0001
  • Test split: 20%
Create Account Login