Face Mask Detection System

AI-Powered Real-Time Mask Compliance Monitoring

** Protect your community with current AI technology**
A comprehensive face mask detection system that combines computer vision, deep learning, and real-time web interfaces to ensure public safety and health compliance.

---

What Makes This Special?

Lightning Fast & Accurate

• Ultra-lightweight models: Only 2.4M parameters for blazing-fast inference
• Real-time detection: Process webcam streams at 30+ FPS
• Multi-face support: Detect and classify multiple faces simultaneously
• High accuracy: 99%+ accuracy on mask detection tasks

Advanced Detection Capabilities

• Correct mask wearing - Green bounding box
• Incorrect mask wearing - Yellow bounding box (nose/mouth exposed)
• No mask detected - Red bounding box with audio alerts

Beautiful Web Interface

• Interactive Dash-based dashboard
• Real-time video streaming
• Image upload and processing
• Configurable confidence thresholds
• Audio alarm system for violations

---

Architecture Overview

graph TB
    A[📷 Input Source] --> B[🔍 Face Detection]
    B --> C[🎭 Mask Classification]
    C --> D[📊 Results Processing]
    D --> E[🖥️ Web Interface]
    D --> F[🔊 Audio Alerts]
    
    B --> G[OpenCV DNN
SSD Face Detector]
    C --> H[MobileNetV2
Mask Classifier]
    E --> I[Dash Web App]
    F --> J[MP3 Audio Files]

---

Quick Start Guide

Option 1: One-Click Setup

# Clone the repository
git clone https://github.com/sidx255/FaceMaskDetection
cd FaceMaskDetection

# Create virtual environment
python -m venv venv

# Activate virtual environment
# Windows:
venv\Scripts\activate
# macOS/Linux:
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the application
python main.py

Option 2: Docker Setup

# Build and run with Docker
docker build -t face-mask-detection .
docker run -p 8050:8050 face-mask-detection

---

How to Use

Web Application

1. Start the server:

   python main.py
   

2. Open your browser: Navigate to http://localhost:8050

3. Choose your mode:

   • Image Mode: Upload and analyze static images
   • Video Mode: Real-time webcam detection

4. Configure settings:

   • Select model (MFN/RMFD)
   • Adjust confidence threshold
   • Enable/disable audio alerts

Command Line Interface

Image Detection

python src/detect_mask_image.py -i example_images/pic1.jpg -m MFN -c 0.5

Video Detection

python src/detect_mask_video.py -m MFN -c 0.5

Model Training

python src/train.py -d MFN

---

AI Models & Performance

Model Comparison

Model	Classes	Parameters	Accuracy	Use Case
MFN	3	2.4M	99.1%	Comprehensive detection
RMFD	2	2.4M	99.3%	Binary classification

Detection Classes

MFN Model (3-Class)

• Correct Mask: Properly worn mask
• Incorrect Mask: Improperly worn (nose/mouth exposed)
• No Mask: No face covering detected

RMFD Model (2-Class)

• With Mask: Face covering detected
• Without Mask: No face covering

---

Technical Stack

Core Technologies

• ** Python 3.7+** - Primary programming language
• ** TensorFlow 2.x** - Deep learning framework
• ** OpenCV 4.x** - Computer vision library
• ** Dash** - Interactive web applications
• ** NumPy** - Numerical computing
• ** Matplotlib** - Data visualization

AI/ML Libraries

• ** MobileNetV2** - Lightweight CNN architecture
• ** OpenCV DNN** - Face detection pipeline
• ** Keras** - High-level neural network API
• ** Scikit-learn** - Machine learning utilities

---

How files are organized

FaceMaskDetection/
├── 🎯 main.py                 # Web application entry point
├── ⚙️ config.py              # Configuration management
├── 📋 requirements.txt        # Python dependencies
├── 🚀 setup.py               # Installation script
├── 📖 README.md              # This file
│
├── 📁 src/                    # Source code
│   ├── 🖼️ detect_mask_image.py    # Image detection
│   ├── 📹 detect_mask_video.py    # Video detection
│   └── 🎓 train.py               # Model training
│
├── 🧠 models/                 # AI models
│   ├── 👤 face_detector/      # Face detection models
│   └── 🎭 mask_detector/      # Mask classification models
│
├── 🎨 assets/                 # Web assets
│   ├── 🎨 base.css            # Styling
│   ├── 🔊 *.mp3               # Audio alerts
│   └── 🖼️ *.jpg               # Example images
│
├── 📊 checkpoints/            # Training checkpoints
├── 📈 figures/                # Training plots
└── 📁 output/                 # Generated outputs

---

Core workflows

Performance Optimized

• Lightweight models for edge deployment
• GPU acceleration support
• Batch processing for multiple faces
• Memory efficient inference pipeline

User Experience

• Intuitive web interface with real-time feedback
• Drag-and-drop image upload
• Responsive design for all devices
• Audio-visual alerts for violations

Developer Friendly

• Modular architecture for easy customization
• Comprehensive documentation and examples
• CLI tools for batch processing
• Training scripts for model fine-tuning

---

Performance Metrics

Model Accuracy

MFN Model Performance:
├── Correct Mask:     99.2% accuracy
├── Incorrect Mask:   98.8% accuracy
└── No Mask:         99.5% accuracy

RMFD Model Performance:
├── With Mask:       99.3% accuracy
└── Without Mask:    99.1% accuracy

Inference Speed

• CPU: ~50ms per frame (Intel i7)
• GPU: ~15ms per frame (NVIDIA GTX 1060)
• Mobile: ~100ms per frame (ARM Cortex-A78)

---

Training Your Own Models

Dataset Preparation

1. Organize your data:

   data/
   ├── MFN/
   │   ├── face_with_mask_correct/
   │   ├── face_with_mask_incorrect/
   │   └── face_no_mask/
   └── RMFD/
       ├── face_with_mask/
       └── face_no_mask/
   

2. Start training:

   python src/train.py -d MFN
   

Training Features

• Data augmentation for reliable models
• Early stopping to prevent overfitting
• Model checkpoints for recovery
• Training visualization with plots

---

Configuration Options

Detection Parameters

# Confidence thresholds
FACE_DETECTION_CONFIDENCE = 0.5    # Face detection threshold
MASK_CLASSIFICATION_CONFIDENCE = 0.5  # Mask classification threshold

# Model selection
MODEL_TYPE = "MFN"  # or "RMFD"

# Audio settings
AUDIO_ALERTS_ENABLED = True
ALERT_VOLUME = 0.8

Web Interface Settings

# Server configuration
HOST = "0.0.0.0"
PORT = 8050
DEBUG = False

# UI customization
THEME = "light"  # or "dark"
LANGUAGE = "en"  # or "es", "fr", etc.

---

Deployment Options

Local Development

python main.py

Production Server

gunicorn main:server -b 0.0.0.0:8050

Docker Deployment

docker run -p 8050:8050 face-mask-detection

Cloud Deployment

• AWS: EC2 with GPU instances
• Google Cloud: AI Platform
• Azure: Machine Learning Service

---

Contributing

We welcome contributions! Here's how you can help:

Bug Reports

• Use GitHub Issues
• Include system information
• Provide error logs

Feature Requests

• Describe the use case
• Explain the expected behavior
• Consider implementation complexity

Code Contributions

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Acknowledgments

• OpenCV team for excellent computer vision tools
• TensorFlow team for the powerful ML framework
• Dash team for the amazing web framework
• Community contributors for feedback and improvements

---

Support & Contact

• ** Issues**: GitHub Issues
• ** Discussions**: GitHub Discussions
• ** Email**: Contact Form

---