ArcFace-YuNet Face Recognition

Project Description

This project is a real-time face recognition system that uses YuNet for face detection and ArcFace for feature embedding extraction.
The system compares the registered face of the main user with faces captured from the camera and calculates similarity using cosine similarity.

Features

• Detecting and cropping faces with YuNet.
• Extracting feature embeddings using ArcFace.
• Normalizing embeddings and calculating cosine similarity for identity verification.
• Real-time face recognition using a webcam.

Personal Responsibility

• Designing and implementing the main user face recognition module using computer vision and ArcFace.
• Preprocessing and cropping faces for ArcFace input.
• Calculating embeddings and implementing cosine similarity for comparison.

Technologies & Libraries

• OpenCV
• NumPy
• InsightFace (ArcFace & YuNet)

Files

• app.py: main project code
• requirements.txt: required Python packages

How to Run

1. Install dependencies:

pip install -r requirements.txt