Real-Time Face Recognition System

A Python application that performs real-time face detection and recognition using a webcam. The system automatically loads known faces from a folder, generates facial encodings, and identifies people in live video streams.

Features

• Real-time face detection and recognition
• Automatic loading of known faces from a directory
• Supports multiple known individuals
• Labels unknown faces as "Unknown"
• Webcam validation and error handling
• Face encoding validation for uploaded images
• Easy to extend by adding new images to the dataset

Technologies Used

• Python 3
• OpenCV
• face_recognition
• NumPy

Project Structure

face-recognition-system/
│
├── face recognize.py
├── requirements.txt
├── README.md
│
└── known_faces/
    ├── Person1.jpg
    ├── Person2.jpg
    └── Person3.jpg

Installation

1. Clone the repository

git clone https://github.com/Matin-python/Face-recognition.git
cd Face-recognition

2. Create a virtual environment (Optional)

python -m venv venv

Activate it:

Windows

venv\Scripts\activate

Linux/macOS

source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

Troubleshooting Installation

The face_recognition library depends on dlib.

First, try installing dlib using:

pip install dlib

If you encounter installation errors while building dlib, install CMake first and then try again:

pip install cmake
pip install dlib

Note: Installing CMake may not be necessary on all systems because precompiled dlib packages are available for many Python versions.

Installing dlib Manually (Windows)

If you are using Windows and still cannot install dlib, you can download a precompiled wheel from the following community-maintained repository:

https://github.com/Murtaza-Saeed/Dlib-Precompiled-Wheels-for-Python-on-Windows-x64-Easy-Installation

Download the wheel file that matches your Python version.

For example, if you are using Python 3.12.x, download:

dlib-19.24.99-cp312-cp312-win_amd64.whl

After downloading the file:

1. Move the .whl file to a folder of your choice.
2. Open Command Prompt in that folder.
3. Run the following command:

pip install dlib-19.24.99-cp312-cp312-win_amd64.whl

Tip: In Windows, you can quickly open Command Prompt in a folder by typing cmd in the File Explorer address bar and pressing Enter.

Note: This repository is maintained by the community and is not an official dlib release. If it becomes unavailable or does not contain a wheel for your Python version, you may need to build dlib from source or look for another compatible precompiled wheel.

NumPy Compatibility

If you encounter the following error:

RuntimeError: Unsupported image type, must be 8bit gray or RGB image.

your installed version of NumPy is likely too new.

Some versions of dlib are not compatible with NumPy 2.x. In that case, install a version of NumPy below 2.0 (see requirements.txt).

Note: Your installed version of OpenCV (cv2) must also be compatible with your NumPy version. Otherwise, the project may fail to run due to dependency conflicts.

Installing face_recognition_models

If you encounter the following message when running the project:

Please install `face_recognition_models` with this command before using `face_recognition`:

install the required models by running:

pip install git+https://github.com/ageitgey/face_recognition_models

Then run the application again.

If the same message still appears after installing face_recognition_models, the issue may be caused by an incompatible version of setuptools.

Try installing an older version of setuptools:

pip install setuptools==80.10.2

After downgrading setuptools, install face_recognition_models again:

pip install git+https://github.com/ageitgey/face_recognition_models

Then run the application again.

Tested Environment

This project has been tested on:

• Windows 11
• Python 3.12
• face_recognition = 1.2.3
• dlib = dlib-19.24.99-cp312-cp312-win_amd64.whl
• opencv-python = 4.10.0.84
• NumPy = 1.26.4

Usage

Add Known Faces

Before running the application, you must add at least one image to the known_faces directory.

Example:

known_faces/
├── Matin.jpg
├── Mahdi.jpg
├── Sara.jpg

The filename (without the extension) will be used as the person's name during recognition.

Important: If the known_faces folder is empty, the system will not be able to recognize anyone.

Run the Application

python face recognize.py

Controls

• Press Q to quit the application.

How It Works

1. Loads all images from the known_faces folder.
2. Generates face encodings for each valid image.
3. Opens the webcam feed.
4. Detects faces in each frame.
5. Compares detected faces against known encodings.
6. Calculates face distances and finds the closest match.
7. Draws a bounding box and displays the detected person's name.

Example Output

Known Face

Matin

Unknown Face

Unknown

Error Handling

The application checks for:

• Missing webcam access
• Invalid image files
• Images without detectable faces
• Empty face database

Example messages:

Cannot access webcam

No face found in Sara.jpg

No known faces found.

Future Improvements

• Face registration through the webcam
• Graphical User Interface (GUI)
• Face recognition confidence scores
• Attendance logging system
• Database integration
• Performance optimization with frame skipping
• Multi-camera support

License

This project is licensed under the MIT License.

Author

Mohammad Reza Bakhshandeh

Electrical Engineering (Electronics) Graduate

Interested in Python Development, Computer Vision, Machine Learning, and Artificial Intelligence.