Face Detection InsightFace

A simple face detection and recognition system using InsightFace for accurate face recognition.

Features

• Real-time face detection from webcam
• Face recognition with known faces
• Confidence scores for recognition accuracy
• Simple and lightweight implementation

Example Images for known_faces (front and side view of fac profile)

Installation

1. Install dependencies:

pip install -r requirements.txt

Usage

1. Add known faces to the known_faces directory:

   • Create subdirectories for each person (e.g., Mario/, Bastien/, etc.)
   • Add clear face images (JPG or PNG format) in each person's directory
   • Face should be clearly visible and well-lit

2. Run the face detection script:

python simple_face_detection.py

Controls

• Press 'q' to quit
• Press 's' to save the current frame

How It Works

1. Face Detection: Uses InsightFace for accurate face detection and recognition
2. Face Recognition: Extracts face embeddings and compares with known faces
3. Real-time Processing: Processes video feed from webcam in real-time

About InsightFace

This project is built using InsightFace, a state-of-the-art 2D&3D deep face analysis toolbox. InsightFace provides:

• Advanced face detection algorithms (RetinaFace, SCRFD)
• High-accuracy face recognition with ArcFace
• Face alignment and landmark detection
• Optimized for both training and deployment

InsightFace is an open-source project that efficiently implements various state-of-the-art algorithms for face recognition, face detection, and face alignment, primarily based on PyTorch and MXNet.

Project Structure

face-detection-insightface/
├── simple_face_detection.py    # Main face detection script
├── known_faces/                # Directory for known face images
│   ├── Mario/                  # Mario's face images
│   ├── Bastien/                # Bastien's face images
│   ├── Freya/                  # Freya's face images
│   └── Shane/                  # Shane's face images
├── requirements.txt            # Python dependencies
└── README.md                   # This file

Requirements

• Python 3.8+
• Webcam or video input device
• Clear, well-lit face images for recognition

License

MIT License