https://github.com/allanotieno254/pneumonia-xray-classification
This project is a Deep Learning-based Pneumonia classification system that allows medical staff to upload chest X-ray images and quickly determine whether a patient shows signs of pneumonia.
https://github.com/allanotieno254/pneumonia-xray-classification
keras machine-learning numpy pandas pil pillow python streamlit streamlit-webapp
Last synced: 3 months ago
JSON representation
This project is a Deep Learning-based Pneumonia classification system that allows medical staff to upload chest X-ray images and quickly determine whether a patient shows signs of pneumonia.
- Host: GitHub
- URL: https://github.com/allanotieno254/pneumonia-xray-classification
- Owner: AllanOtieno254
- Created: 2025-08-23T08:12:04.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-08-23T17:14:24.000Z (11 months ago)
- Last Synced: 2025-10-11T05:19:40.410Z (9 months ago)
- Topics: keras, machine-learning, numpy, pandas, pil, pillow, python, streamlit, streamlit-webapp
- Language: Python
- Homepage:
- Size: 2.13 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Pneumonia Classification Using Chest X-ray Images
Dataset:https://data.mendeley.com/datasets/rscbjbr9sj/2
## Table of Contents
- [Project Overview](#project-overview)
- [Features](#features)
- [Project Structure](#project-structure)
- [Installation](#installation)
- [Usage](#usage)
- [Model Details](#model-details)
- [Integration into Hospital Systems](#integration-into-hospital-systems)
- [Benefits and Usefulness](#benefits-and-usefulness)
- [License](#license)
---
## Project Overview
This project is a **Deep Learning-based Pneumonia classification system** that allows medical staff to upload chest X-ray images and quickly determine whether a patient shows signs of pneumonia. The system uses a pre-trained Keras model to predict the condition and provides a confidence score for the diagnosis.
The project is built using:
- **Python**
- **Streamlit** for web interface
- **Keras / TensorFlow** for model inference
- **Pillow (PIL)** and **NumPy** for image processing
This tool is designed for **rapid, accurate preliminary screening** in medical facilities, reducing the burden on radiologists and improving response time for patient care.
---
## Features
- Upload a chest X-ray image in multiple formats (`jpeg`, `jpg`, `png`, `dcm`, `tiff`, `bmp`, `gif`)
- Automatic classification of **Pneumonia** (label 0) or **Normal** (label 1)
- Confidence score displayed for each prediction
- Simple, interactive web interface using Streamlit
- Preprocessing steps for consistent image input (resizing, normalization)
- Option to integrate a custom background for branding
---
## Project Structure
```
pneumonia-classification/
│
├── main.py # Streamlit application
├── util.py # Utility functions: classify images, set background
├── model/
│ ├── keras_model.h5 # Pre-trained pneumonia classification model
│ └── labels.txt # Labels: Pneumonia (0), Normal (1)
├── requirements.txt # Python dependencies
└── README.md # Project documentation
```
---
## Installation
1. **Clone the repository**
```bash
git clone https://github.com/YourUsername/pneumonia-classification.git
cd pneumonia-classification
```
2. **Create and activate a virtual environment**
```bash
python -m venv venv
source venv/bin/activate # Linux / macOS
venv\Scripts\activate # Windows
```
3. **Install dependencies**
```bash
pip install -r requirements.txt
```
4. **Run the Streamlit app**
```bash
streamlit run main.py
```
---
## Usage
1. Open the Streamlit web app.
2. Upload a chest X-ray image using the file uploader.
3. The app will display the uploaded image.
4. Classification result will appear below the image along with the confidence score.
> Example Output:
```
Pneumonia
Score: 0.98
```
---
## Model Details
- **Input Size:** 224x224 pixels
- **Normalization:** [-1, 1] scale
- **Custom Thresholding:** Images are labeled Pneumonia if probability > 0.95, else Normal
- **Output:** Class name (`Pneumonia` or `Normal`) and confidence score
This approach ensures **high accuracy for pneumonia detection** while minimizing false negatives for normal cases.
---
## Integration into Hospital Systems
The system can be integrated into hospital workflows in several ways:
1. **Electronic Medical Records (EMR) Integration:**
- The app can automatically save classification results to patient records for radiologists to review.
2. **Triage Support:**
- Patients showing high-confidence Pneumonia predictions can be prioritized for urgent care.
3. **Remote Diagnostics:**
- Hospitals in rural areas can upload X-rays to this system and receive preliminary screening results before consulting specialists.
4. **Dashboard and Reporting:**
- Aggregate statistics can be generated for hospital administration to monitor pneumonia cases and trends over time.
---
## Benefits and Usefulness
- **Speed:** Provides near-instantaneous preliminary results.
- **Accuracy:** Reduces human error in X-ray interpretation.
- **Accessibility:** Can be used in hospitals without enough radiologists.
- **Cost-effective:** Requires only a standard computer and X-ray images; reduces unnecessary tests.
- **Data-driven decisions:** Helps hospitals track outbreaks, patient load, and optimize resources.
---
## License
This project is licensed under the **MIT License**. See `LICENSE` file for details.