🧠 Brain Tumor Detection with CNN + Gradio

This project implements a deep learning model using Convolutional Neural Networks (CNN) to detect brain tumors from MRI scans. It features a clean, interactive Gradio interface that allows users to upload an image and receive a classification: Tumor or No Tumor.

---

📁 Dataset Used

Brain MRI Images for Brain Tumor Detection

• Binary classes: yes (tumor present) and no (no tumor)
• Downloaded using kagglehub within the notebook

---

🏗️ Model Architecture

Custom CNN built with TensorFlow/Keras:

• Input: 128x128 RGB image
• Layers:
  • Conv2D (32 filters) → MaxPooling
  • Conv2D (64 filters) → MaxPooling
  • Flatten → Dropout (0.5)
  • Dense (64) → Dense (2 - softmax)

---

🔍 What the Project Does

• Downloads and loads MRI dataset using kagglehub
• Preprocesses and normalizes images
• Trains a CNN model to classify scans as Tumor/No Tumor
• Provides an interactive Gradio interface for image-based prediction

---

🔧 Tech Stack

Tech stack: Python, TensorFlow/Keras, NumPy, OpenCV, Gradio, Google Colab, Kaggle API

---

🚀 Google Colab Notebook

👉 Try the Colab Notebook

---

🖼 Demo

---

📌 Future Enhancements

• Add Grad-CAM for explainability
• Include multi-class tumor classification
• Deploy model as a Hugging Face Space

---

📜 License

MIT License (or your choice)