🎭 ArLface Recognition System

Advanced Face Recognition Platform powered by ArcFace (AuraFace)

A modern face recognition system featuring real-time webcam tracking, multi-face detection, video processing, face comparison, and a premium glassmorphism UI. Built for research, learning, and advanced experimentation.

---

📸 Application Preview

Base

Train

https://drive.google.com/file/d/1h_xa7KNIEb8MYYG819ct6I4MrC8od3_g/view?usp=sharing

Test

https://drive.google.com/file/d/1L-_VyJx2jcJoe20XZLq0ZivNL_vfEGiz/view?usp=sharing https://drive.google.com/file/d/1_CRLqrGKArtgvbKWolrgJo2Uh_K7H7OU/view?usp=sharing

Compare

https://drive.google.com/file/d/199K7igRlfHVtWFJcpxZHVqpjH3mBZL82/view?usp=sharing

Live

Analytics

Settings

---

🤖 AI Model

Model Name: AuraFace (ArcFace)

🔗 https://huggingface.co/fal/AuraFace-v1

⚠️ Model Usage Disclaimer

This project does not develop, modify, or claim ownership of any face recognition model.

AuraFace is used strictly as a pretrained backbone for:

• Face embedding extraction
• Similarity matching

All system logic, pipelines, UI, backend, and application behavior are fully custom-built by the author.

---

✨ Features

🎯 Core Capabilities

• Real-time face recognition (webcam)
• Multi-face detection (10+ faces)
• Photo & video processing
• Face comparison & similarity scoring
• Age & gender estimation
• 106 facial landmarks detection
• UNKNOWN face handling

⚙️ Advanced Features

• Configurable similarity threshold
• Multiple similarity metrics
• Video frame auto-extraction
• Export / import database
• Real-time analytics dashboard

🎨 User Interface

• Glassmorphism dark UI
• Purple → Cyan gradient theme
• Fully responsive layout
• Smooth animations

---

🚀 Installation

Prerequisites

• Python 3.8+
• pip
• Webcam (for Live mode)

git clone https://github.com/Arlchoose-code/ArLface-Recognition.git
cd ArLface-Recognition
pip install -r requirements.txt
python main.py

Open in browser:

http://localhost:8000

All AI models are already included in the repository.

---

📖 Usage Guide

1️⃣ Train

• Add new or existing person
• Upload images or videos
• Videos auto-extract 20 frames

2️⃣ Test

• Upload images or videos
• Display name, confidence, age & gender

3️⃣ Live

• Real-time webcam recognition

4️⃣ Compare

• Compare two faces
• Adjustable similarity threshold

5️⃣ Analytics

• Recognition statistics
• Performance insights

6️⃣ Settings

• Threshold & metrics
• Database management

---

🧱 Tech Stack

Backend

• FastAPI
• InsightFace
• OpenCV
• NumPy

AI Models

• ArcFace (AuraFace)
• RetinaFace / SCRFD
• 2D106det
• GendAge

Frontend

• TailwindCSS
• Vanilla JavaScript
• HTML5 Canvas

Storage

• JSON
• Pickle

---

📁 Project Structure

ArLface-Recognition/
├── main.py
├── requirements.txt
├── README.md
├── app/
│   ├── database.py
│   ├── face_detector.py
│   └── recognizer.py
├── models/
│   └── auraface/
├── templates/
└── data/

---

🔒 Privacy & Security

• Local processing only
• No cloud & no telemetry
• Full data ownership

---

📄 License & Commercial Use

Released under the MIT License.

✅ Allowed

• Personal & educational use
• Research & experimentation

⚠️ Commercial Use

Commercial usage requires permission from:

1. AuraFace (model license)
2. Project author

Permission is free, confirmation only.

---

👨‍💻 Author

Syahril Haryono GitHub: https://github.com/Arlchoose-code Instagram: @syahrilh.h

---

🙏 Acknowledgments

• AuraFace
• ArcFace
• InsightFace
• FastAPI
• TailwindCSS

---

⭐ Star this repository if you find it useful Made with ♥ by Syahril Haryono