COVID-19 Face Mask Detector

A real-time face mask detection system using deep learning and computer vision to identify whether people are wearing face masks. Built with TensorFlow, OpenCV, and Streamlit.

Features

• Real-time Detection: Detect face masks in uploaded images with bounding boxes and confidence scores
• User-Friendly Interface: Interactive Streamlit web application for easy image uploads
• High Accuracy: Uses MobileNetV2 architecture for efficient and accurate predictions
• Visual Feedback: Color-coded bounding boxes (Green for mask, Red for no mask)
• Confidence Scores: Displays prediction confidence percentage for each detection

Tech Stack

• Deep Learning: TensorFlow/Keras with MobileNetV2
• Computer Vision: OpenCV for image processing and face detection
• Frontend: Streamlit for the web interface
• Model: Pre-trained face detector with custom mask classification model

Installation

Prerequisites

• Python 3.8 or higher
• pip package manager

Setup

1. Clone the repository:

git clone https://github.com/CodeBalch25/Face_Mask_Detection_Python.git
cd Face_Mask_Detection_Python

2. Install dependencies:

pip install -r requirements.txt

3. Download the required model files:

   • Place face_mask_detector.model in the root directory
   • Place face detector files in face_detector/ directory:
     • deploy.prototxt
     • res10_300x300_ssd_iter_140000.caffemodel

4. Create the necessary directories:

mkdir images
mkdir face_detector

Usage

Run the Streamlit application:

streamlit run app.py

The application will open in your browser at http://localhost:8501

Using the App

1. Select "Image" from the sidebar
2. Upload an image (JPG or PNG format)
3. Click "Process" to detect face masks
4. View the results with bounding boxes and confidence scores

Project Structure

Face_Mask_Detection_Python/
├── app.py                          # Main Streamlit application
├── detect_mask_in_image.py         # Image detection module
├── detect_mask_in_video.py         # Video detection module (future feature)
├── searching_face.py               # Face search utilities
├── training_detection.py           # Model training script
├── requirements.txt                # Python dependencies
├── face_detector/                  # Pre-trained face detector models
│   ├── deploy.prototxt
│   └── res10_300x300_ssd_iter_140000.caffemodel
├── images/                         # Temporary image storage
└── README.md                       # Project documentation

How It Works

1. Face Detection: Uses OpenCV's DNN module with a pre-trained Caffe model to detect faces
2. Preprocessing: Detected faces are extracted, resized to 224x224, and preprocessed for the neural network
3. Classification: MobileNetV2-based model classifies each face as "Mask" or "No Mask"
4. Visualization: Draws bounding boxes with labels and confidence scores on the original image

Model Architecture

• Base Model: MobileNetV2 (pre-trained on ImageNet)
• Input Size: 224x224x3
• Output: Binary classification (Mask / No Mask)
• Face Detector: SSD (Single Shot Detector) with ResNet-10 backbone

Future Enhancements

• [ ] Real-time webcam detection
• Batch processing for multiple images
• Mobile app deployment
• Dataset expansion for improved accuracy
• Support for different mask types

Technical Details

Detection Pipeline:

1. Image input → Blob conversion
2. Face detection using DNN
3. ROI extraction and preprocessing
4. Mask classification
5. Result visualization

Performance:

• Detection confidence threshold: 0.5
• Image preprocessing: BGR to RGB conversion
• Real-time capable on modern hardware

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is open source and available under the MIT License.

Author

Timothy Balch - @CodeBalch25

Acknowledgments

• TensorFlow team for the MobileNetV2 architecture
• OpenCV community for computer vision tools
• Streamlit for the excellent web framework

Tags

machine-learning computer-vision tensorflow opencv covid19 deep-learning streamlit python face-detection image-classification