Table of Contents

• Table of Contents
• Revolutionizing Attendance with Edge Computing and Deep Learning
• Features
• Hardware Requirements
• Software Dependencies
• Installation
• Usage
  • 1. Encode Face Database
  • 2. Real-time Recognition
  • 3. Batch Processing
• Project Structure
• Configuration
• Contributing
• License
• Acknowledgments
• Presentation

Revolutionizing Attendance with Edge Computing and Deep Learning

Welcome to PiPresence—an automated attendance tracking solution powered by YOLOv8 and MobileFaceNet, designed to run efficiently on edge devices like the Raspberry Pi 4B. By leveraging state-of-the-art deep learning models and edge computing, PiPresence delivers accurate, real-time face detection and recognition without requiring external servers.

Features

• Advanced Face Detection: Uses YOLOv8n for efficient and accurate face detection
• Robust Face Recognition: Implements MobileFaceNet for reliable face recognition
• Multiple Profile Support: Handles left, front, and right face profiles for better accuracy
• Edge Processing: All computations performed locally on the Raspberry Pi
• Real-time Processing: Supports both real-time camera feed and batch image processing
• Flexible Integration: Choice of weighted average, clustering, or distance-based methods for face embedding
• Comprehensive Logging: Detailed logging system with configurable verbosity

Hardware Requirements

• Raspberry Pi 4B
• Raspberry Pi Camera Module
• MicroSD Card (16GB or larger recommended)
• Power Supply for Raspberry Pi
• Optional: Monitor, Keyboard, and Mouse for setup

Software Dependencies

• Python 3.9 or higher
• Poetry for dependency management
• Key libraries (automatically managed by Poetry):
  • OpenCV Python
  • ONNX Runtime
  • NumPy
  • FAISS-CPU
  • Click
  • Protobuf

Installation

1. Clone the Repository

git clone <repository_url>
cd pipresence

2. Install Poetry (if not already installed)

curl -sSL https://install.python-poetry.org | python3 -

3. Install Dependencies

poetry install

4. Set Up Environment

# Activate the virtual environment
poetry shell

Usage

PiPresence offers several operation modes:

1. Encode Face Database

Process and encode faces from input images:

pipresence --encode --input-dir /path/to/raw/images --output-dir /path/to/processed/faces

2. Real-time Recognition

Run real-time face recognition using camera feed:

pipresence --infer --camera --verbose

3. Batch Processing

Process a directory of images:

pipresence --infer --input-dir /path/to/images --output-dir /path/to/output

Project Structure

pipresence/
├── data/
│   ├── models/              # Pre-trained models
│   │   ├── yolov8n-face.onnx
│   │   └── mobilefacenet_fixed.onnx
│   ├── encodings/           # Face embeddings database
│   └── images/             # Input images
├── pipresence/
│   ├── detect_faces.py     # Face detection using YOLOv8
│   ├── recognize_faces.py  # Face recognition using MobileFaceNet
│   ├── preprocess.py      # Image preprocessing
│   ├── config.py          # Configuration settings
│   ├── main.py           # Main application logic
│   └── tools/
│       └── utils.py      # Utility functions
└── pyproject.toml        # Project dependencies and metadata

Configuration

Key parameters can be modified in config.py:

• Detection and recognition thresholds
• Image preprocessing settings
• Model paths
• Directory paths
• Logging verbosity

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests (if available)
5. Submit a pull request

License

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

Acknowledgments

• YOLOv8 for efficient face detection
• MobileFaceNet for accurate face recognition
• ONNX Runtime for optimized model inference
• FAISS for efficient similarity search

Presentation

The presentation for PiPresence is available in the Assets folder under /doc. You can view it via the following link:

View PiPresence Presentation

---

For issues, feature requests, or contributions, please open an issue or pull request on the repository.