README.md

Brain Tumor Classifier

This repository contains project on Brain Tumor Classification.

Task

Given MRI scans of the brain, the goal is to identify the type of brain tumor.

The images belong to the following 4 classes:

1. Glioma Tumor
2. Meningioma Tumor
3. No Tumor
4. Pituitary Tumor

In essence, this is an image classification problem.

Samples from Tumor Classes

Approach

EfficientNet is an efficent model that reaches State-of-the-Art accuracy on image classification transfer learning tasks. In this problem, EfficientNet with weights pre-trained on imagenet are used.

Model training curves

The performance metrics of the model with respect to number of epochs can be visualised as follows:

Result

The model yielded > 95% accuracy when tested a large number of times.