🧠 Brain Tumor Classification using Deep Learning (Custom CNN)

📌 Project Overview

This project focuses on Brain Tumor Classification from MRI images using a Custom Convolutional Neural Network (CNN) built from scratch with TensorFlow and Keras.

The model classifies MRI scans into four categories:

• Glioma
• Meningioma
• Pituitary Tumor
• No Tumor

The goal of this project is to build an accurate, well-generalized, and explainable deep learning model and deploy it using Streamlit for real-time predictions.

---

🎯 Key Highlights

• ✅ Custom CNN architecture (no transfer learning)
• ✅ Proper train–validation–test split
• ✅ Data augmentation tuned for medical MRI images
• ✅ Learning rate scheduling and early stopping
• ✅ Achieved 85.2% accuracy on unseen test data
• ✅ Deployed using Streamlit with live predictions

---

📂 Dataset Information

• Source: Kaggle – Brain Tumor MRI Dataset
• Image Type: MRI scans (visually grayscale, stored as RGB)
• Classes: 4

Dataset Structure

Dataset/
├── Training/
│   ├── glioma/
│   ├── meningioma/
│   ├── pituitary/
│   └── notumor/
└── Testing/
    ├── glioma/
    ├── meningioma/
    ├── pituitary/
    └── notumor/

Dataset Split Strategy

• Training data: Used for model learning
• Validation data: 20% split from training set
• Test data: Completely unseen during training

---

🛠️ Technologies Used

• Python
• TensorFlow / Keras
• NumPy
• Matplotlib
• Google Colab
• Streamlit

---

🧠 Model Architecture (Custom CNN)

The model was designed from scratch using convolutional blocks optimized for medical images:

• Convolution + ReLU activation
• Batch Normalization
• MaxPooling layers
• Dropout for regularization
• Dense layers for classification
• Softmax output layer

Input Shape

(224, 224, 3)

MRI images appear grayscale but are stored and processed as RGB images with identical channels.

---

🔁 Data Preprocessing & Augmentation

MRI-safe augmentation techniques were applied:

• Image resizing and normalization
• Small rotations
• Zooming
• Width and height shifting
• Horizontal flipping

Aggressive augmentation was avoided to preserve medical image integrity.

---

⚙️ Training Strategy

• Optimizer: Adam

• Learning Rate: 3e-4

• Loss Function: Categorical Crossentropy

• Callbacks:

  • EarlyStopping (restore best weights)
  • ReduceLROnPlateau

The best model was automatically restored based on validation loss.

---

📈 Model Performance

Metric	Value
Training Accuracy	~92%
Best Validation Accuracy	82.3%
Final Test Accuracy	85.2%
Test Loss	0.6078

---

🖼️ Training Results Visualization

The following plot shows training and validation accuracy during model training:

---

🌐 Streamlit Deployment

The trained model is deployed using Streamlit, providing an interactive interface for real-time MRI classification.

Application Features

• Upload MRI images
• Real-time tumor classification
• Clear display of predicted class

---

🎥 Streamlit Application Demo Video

Click the link below to watch the Streamlit app running live:

▶️ Watch Streamlit Demo Video

ℹ️ GitHub does not play videos inline. The link is clickable and downloadable.

---

💾 Model Saving

The trained model is saved in both formats:

brain_tumor_model.h5
brain_tumor_model.keras

• .h5 → Deployment & compatibility
• .keras → Future-proof format

---

🧪 How to Run the Project

1️⃣ Clone the Repository

git clone https://github.com/Sayeem3051/MRI-Classification1.git
cd MRI-Classification1

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Run Streamlit App

streamlit run app.py

---

🧠 Learning Outcomes

• Built a CNN from scratch for medical imaging
• Applied correct validation strategies
• Controlled overfitting using callbacks
• Deployed a deep learning model using Streamlit

---

🚀 Future Improvements

• Use transfer learning (MobileNetV2 / ResNet)
• Add Grad-CAM for model explainability
• Improve class-wise performance
• Deploy as a public web application

---

🏁 Conclusion

This project demonstrates a complete deep learning pipeline from data preprocessing and model training to evaluation and deployment. The achieved performance validates the effectiveness of a well-designed Custom CNN for brain tumor classification.

---

⭐ If you find this project useful, consider starring the repository!