🧠 Real-Time Face Recognition System

This project is a Python-based face recognition system that detects faces from images, encodes them, and stores the embeddings in a pickle file (embeddings.pkl). It can dynamically update with new faces and supports real-time identification.

---

📂 Project Structure

├── dataset/
│ ├── Person1/
│ │ ├── img1.jpg
│ │ └── img2.jpg
│ ├── Person2/
│ │ └── img1.jpg
├──  collectingDatasets.py
├── embeddings.pkl
├── generate_embeddings.py
├── face_recognition_live.py
├── requirements.txt
└── README.md

---

🔧 Features

• Generate face encodings using face_recognition
• Automatically update embeddings.pkl if it already exists
• Warns if no face is found in an image
• Real-time face recognition support (optional)
• Error-handling and logging for easy debugging

---

🛠️ Installation

1. Clone the repository

git clone https://github.com/your-username/face-recognition-system.git
cd face-recognition-system

2. Create a virtual environment (optional but recommended)

python -m venv venv
venv\Scripts\activate  # On Windows
source venv/bin/activate  # On Linux/Mac

3. Install dependencies

pip install -r requirements.txt

▶️ Usage

To generate/update face embeddings:

python generate_embeddings.py

If embeddings.pkl already exists, new embeddings will be added without duplicating existing ones.

To recognize faces (optional):

python recognize_faces.py

🗂️ Dataset Format

Your dataset/ folder should be structured like:

dataset/
├── Alice/
│   ├── alice_1.jpg
│   └── alice_2.jpg
├── Bob/
│   ├── bob_1.jpg

Each subfolder represents a different person. Images should have clear frontal faces.

📦 Requirements

Python 3.8+

face_recognition

numpy

Pillow

dlib (C++ build tools required)

📌 Note: Installing dlib requires Microsoft C++ Build Tools on Windows.

📄 License

This project is licensed under the MIT License.

🙋‍♂️ Author

Divyansu Giri