Face Recognition System

A high-performance Face Recognition System built with Python, OpenCV, and Deep Learning. This system is designed to identify and verify individuals in real-time using convolutional neural networks (CNNs), with a strong emphasis on data security.

Key Features

• Deep Learning Integration: Utilizes state-of-the-art CNN-based face detection and 128-dimensional feature extraction for high accuracy.
• Real-time & Static Recognition: Processes live video streams via webcam or verifies identities from stored static images.
• Encrypted Storage: Ensures absolute data privacy by encrypting facial embeddings using symmetric AES encryption (Fernet) before storing them in the local SQLite database.
• Threshold-based Verification: Employs mathematical Euclidean distance thresholding to prevent false positives and verify identities securely.
• Cyberpunk HUD: Features a custom-built, futuristic OpenCV targeting interface with live FPS telemetry.

Tech Stack

• Python 3
• OpenCV (opencv-python): Real-time computer vision and HUD rendering.
• Face_Recognition / Dlib: CNN feature extraction.
• Cryptography: Encrypted biometric storage.
• SQLite: Local database management.

Installation

1. Clone the repository:

   git clone https://github.com/2k33cse992574/face_detection_python_app.git
   cd face_detection_python_app

2. Create a virtual environment and install dependencies:

   python -m venv venv
   # Windows
   .\venv\Scripts\activate
   # Linux/Mac
   source venv/bin/activate
   
   pip install -r requirements.txt

   (Note: Installing on Windows may require Visual Studio C++ Build Tools or a precompiled dlib wheel).

Usage

1. Register a User

To register your face into the encrypted database:

python main.py --register "Your Name"

(Press 's' to scan your face and save securely).

2. Live Webcam Verification

To start the real-time recognition stream:

python main.py --recognize

3. Static Image Verification

To run recognition against a stored photo:

python main.py --image path/to/photo.jpg

4. Enable CNN Mode (Optional)

For maximum accuracy (recommended if you have a GPU), append the --cnn flag:

python main.py --recognize --cnn

Security Note

This repository is configured to strictly ignore the secret.key and face_data.db files. Never commit your secret.key to version control, as it is required to decrypt the biometric facial embeddings.