Face Recognition with Liveness Detection

An automated employee attendance system using face recognition with anti-spoofing capabilities for my previous company. This system eliminates the need for fingerprint scanners or manual attendance marking by providing secure, real-time face authentication with liveness detection.

Overview

This project implements a comprehensive face recognition system that includes:

Real-time face detection and recognition
Liveness detection to prevent spoofing attacks
Pose and emotion detection capabilities
Automated attendance tracking
Web-based interface using Flask

Requirements

Python: 3.6+ (Originally developed with Python 3.6/3.7)
Operating System: Windows, macOS, or Linux
Hardware: Webcam, GPU recommended for real-time processing
Dependencies: Listed in requirements.txt

Features

Face Recognition: Accurate employee identification using trained facial models
Anti-Spoofing: Liveness detection to prevent photo/video spoofing attempts
Real-time Processing: Live webcam-based attendance marking
Pose Detection: Employee pose analysis for better authentication
Emotion Recognition: Optional emotion detection capabilities
Web Interface: Flask-based web application for easy deployment
Automated Training: Tools for training custom face recognition models

Project Structure

face_recog_livliness/
├── datasets/                           # Training datasets and employee photos
├── libfaceid/                         # Core face recognition library
├── models/                            # Trained ML models and weights
├── templates/                         # HTML templates for web interface
├── final_basic_real_time_face_recog.py    # Basic real-time face recognition
├── final_fr_with_liveliness.py           # Face recognition with liveness detection
├── final_real_time_pose_emotion.py       # Pose and emotion detection
├── final_training_model.py               # Model training script
├── training.py                           # Training utilities
├── xahid_training.py                     # Custom training implementation
├── testing_webcam.py                     # Webcam testing utilities
├── testing_webcam_flask.py              # Flask web interface testing
├── testing_webcam_livenessdetection.py  # Liveness detection testing
├── xahid_blinking.py                     # Blink detection for liveness
├── xahid_demo.py                         # Demo implementation
└── requirements.txt                      # Python dependencies

Installation

Prerequisites

Python 3.6 or higher (Python 3.7+ recommended)
pip package manager
Webcam or camera device

Setup Instructions

Clone the repository:

git clone https://github.com/AnwarXahid/face_recog_livliness.git
cd face_recog_livliness

Create virtual environment (recommended):

# For Windows
python -m venv face_recog_env
face_recog_env\Scripts\activate

# For macOS/Linux
python3 -m venv face_recog_env
source face_recog_env/bin/activate

Install dependencies:
```
pip install -r requirements.txt
```
Set up datasets:
- Place employee photos in the datasets/ directory
- Organize photos by employee ID or name
- Ensure good quality, well-lit photos for better recognition

Usage

Training the Model

Prepare your dataset:
```
python training.py
```
Train the face recognition model:
```
python final_training_model.py
```
Custom training (alternative):
```
python xahid_training.py
```

Running the System

Basic Face Recognition:

python final_basic_real_time_face_recog.py

Face Recognition with Liveness Detection:
```
python final_fr_with_liveliness.py
```
With Pose and Emotion Detection:
```
python final_real_time_pose_emotion.py
```
Web Interface:
```
python testing_webcam_flask.py
```

Testing Components

Test webcam functionality:
```
python testing_webcam.py
```

Test liveness detection:

python testing_webcam_livenessdetection.py

Test blink detection:
```
python xahid_blinking.py
```
Run demo:
```
python xahid_demo.py
```

Key Components

Liveness Detection

The system implements multiple liveness detection methods:

Blink Detection: Monitors eye blinks to ensure a live person
Face Movement: Tracks subtle facial movements
Depth Analysis: Analyzes facial depth to prevent photo spoofing