KAUSSHIK / BrainTumorClassification

An attempt to use CNNs to classify Brain Tumors as Glioma , No Tumor, Meningioma, Pituitary

Brain Tumor Classification with Convolutional Neural Networks

This repository contains code for a brain tumor classification project using Convolutional Neural Networks (CNNs) with TensorFlow and Keras. The project aims to classify brain tumors in MRI images into distinct categories.

Project Overview

• Objective: To develop a CNN model capable of accurately categorizing brain tumors in MRI images.
• Tools and Technologies: TensorFlow, Keras, Python, IPython Widgets, OpenCV
• Dataset: The model was trained and tested on a dataset comprising over 3,000 MRI images.

Key Features

• Implemented a CNN architecture using transfer learning with EfficientNet pre-trained on ImageNet.
• Achieved an impressive accuracy of 97% on the test set.
• Developed an interactive UI using IPython widgets for real-time image classification.
• Conducted thorough validation using real-world medical data, ensuring practical applicability.