Skip to content

README.md

Latest commit

 

History

History
214 lines (142 loc) · 4.41 KB

README.md

File metadata and controls

214 lines (142 loc) · 4.41 KB
Raw

🫁 Pneumonia Detection from Chest X-ray

An end-to-end deep learning application for detecting Pneumonia from chest X-ray images using a fine-tuned ResNet18 model. The project demonstrates the complete machine learning workflow—from data preprocessing and model training to evaluation and deployment through a Streamlit web application.

📌 Overview

Pneumonia is a potentially life-threatening lung infection where early diagnosis is critical. This project leverages transfer learning to classify chest X-ray images into one of two categories:

  • ✅ Normal
  • 🦠 Pneumonia

The primary objective is to build an accurate and user-friendly diagnostic assistance tool while showcasing an end-to-end machine learning pipeline.

✨ Features

  • Deep learning–based binary image classification
  • Transfer learning with pretrained ResNet18
  • Interactive web interface built with Streamlit
  • Real-time inference on uploaded X-ray images
  • Clean and modular project structure
  • Easy-to-understand training and evaluation scripts

🧠 Model Architecture

  • Base Model: ResNet18 (ImageNet pretrained)
  • Framework: PyTorch
  • Transfer Learning: Fine-tuned final classification layer
  • Output Classes: Normal, Pneumonia
  • Loss Function: CrossEntropyLoss
  • Optimizer: Adam

📊 Model Performance

Metric Score
Accuracy ~95%
Pneumonia Recall ~96%
False Negatives 32

Note: In medical image classification, Recall is often more important than overall accuracy because missing a pneumonia case (false negative) can have serious clinical consequences.

🖥️ Application Preview

Application Screenshot

📁 Project Structure

pneumonia-detection/
│
├── app.py                  # Streamlit application
├── src/
│   ├── data_loader.py
│   ├── model.py
│   ├── train.py
│   └── evaluate.py
│
├── requirements.txt
├── README.md
└── screenshot.png

⚙️ Installation

Clone the repository:

git clone https://github.com/Abhhiiissshhek/pneumonia-detection.git
cd pneumonia-detection

Create a virtual environment (recommended):

python -m venv .venv

Activate the environment:

Windows

.venv\Scripts\activate

Linux / macOS

source .venv/bin/activate

Install dependencies:

pip install -r requirements.txt

Run the application:

streamlit run app.py

📂 Dataset

This project uses the Chest X-ray Pneumonia Dataset available on Kaggle.

The dataset is not included in this repository due to its size and licensing restrictions.

🚀 Training

To train the model from scratch or fine-tune the network:

python src/train.py

After training, save the model weights and use them with the Streamlit application for inference.

📈 Evaluation

Model evaluation includes:

  • Accuracy
  • Recall
  • Confusion Matrix
  • Classification Metrics

The evaluation script can be executed using:

python src/evaluate.py

⚠️ Repository Notes

  • The trained .pth model file is not included because it exceeds GitHub's file size limit.
  • You can reproduce the model by training it using the provided scripts.

🔮 Future Improvements

  • Confidence score for predictions
  • Grad-CAM visualizations for model explainability
  • Docker support
  • Cloud deployment
  • Model comparison (ResNet, EfficientNet, DenseNet)
  • Improved recall with advanced augmentation techniques
  • Continuous Integration (GitHub Actions)

🛠️ Tech Stack

  • Python
  • PyTorch
  • TorchVision
  • Streamlit
  • NumPy
  • Pillow
  • Matplotlib

📚 Key Learnings

Through this project, I gained hands-on experience with:

  • Transfer Learning
  • Medical Image Classification
  • Deep Learning using PyTorch
  • Model Evaluation and Performance Analysis
  • Building production-ready ML applications with Streamlit
  • End-to-end machine learning workflows

👨‍💻 Author

Abhishek Prajapati

⭐ Support

If you found this project useful, consider giving the repository a star. Feedback, suggestions, and contributions are always welcome.