Brain Tumor Classification

Advanced deep learning system for accurate classification of brain tumors from MRI scans using Convolutional Neural Networks

Classify Brain MRI

Upload a brain MRI scan image to classify the type of tumor present, or try with a random sample.

Upload MRI Image

Drag & drop your file here or click to browse

Supports: JPG, PNG, DICOM

MRI Preview

Processing image...

Batch Processing

Upload multiple MRI scans for batch classification

Drag & drop multiple files or click to browse

Classification Results

-
Confidence: -

Detailed Analysis

Model Explanation

Heatmap showing regions that influenced the classification decision

Grad-CAM Heatmap

About Our Model

Our brain tumor classification system uses a Convolutional Neural Network (CNN) trained on MRI scans to identify four different types of brain conditions.

Glioma

Gliomas are tumors that occur in the brain and spinal cord. They begin in the glial cells that surround and support nerve cells.

Meningioma

Meningiomas are tumors that arise from the meninges — the membranes that surround your brain and spinal cord.

No Tumor

Healthy brain tissue with no detectable tumor present in the MRI scan.

Pituitary

Pituitary tumors are abnormal growths that develop in the pituitary gland, which is located at the base of the brain.

How It Works

Our system uses a deep learning model trained on thousands of MRI scans to accurately classify brain tumors.

CNN Architecture

Architecture
Training Process
Performance

Model Architecture

Our brain tumor classification system uses a Convolutional Neural Network (CNN) with multiple convolutional layers, batch normalization, and dropout for regularization. The model was trained on a dataset of brain MRI scans to classify tumors into four categories.

4 Convolutional Layers
1,000+ MRI Samples
4 Classification Categories

Training Process

The model was trained using the Adam optimizer with a learning rate of 0.0005. We used categorical cross-entropy as the loss function and implemented data augmentation techniques to improve generalization.

Data Augmentation

  • Random rotations (±15°)
  • Random horizontal & vertical flips
  • Brightness & contrast adjustments
  • Zoom range: 0.9 to 1.1

Training Parameters

  • Batch size: 32
  • Epochs: 50 with early stopping
  • Validation split: 20%
  • Test split: 10%

Model Performance

Our model achieves state-of-the-art performance on brain tumor classification tasks, with high accuracy across all tumor types.

96.8%
Accuracy
95.7%
Precision
95.2%
Recall

Our Team

Yash Naidu

Backend & Model Development

Malkapuri Rahul Baswaraj

Frontend & Data Analysis

Dr. Suma Kamalesh Gandhimathi

Faculty Guide