Manual roll calls
Slow for large classes and disruptive during lectures, labs, and tutorials
AI Smart Campus
Facial Recognition Attendance System
A modern attendance platform that uses facial recognition to verify students, automate classroom records, and help lecturers reduce manual work with a faster, cleaner and more secure process
Live session
39 verified 3 pending matchesBackground
Attendance should feel invisible
Traditional attendance methods can create avoidable admin work and leave room for inaccurate records
Bluetooth check-ins
Dependent on student devices, signal reliability, and stable connectivity
Card-based systems
Cards can be lost, shared, or used for proxy attendance without identity verification
AI Recognition Pipeline
SCRFD + ArcFace
Solution
Secure attendance, captured automatically.
VisionAttend uses AI-powered face detection and recognition to identify students in classroom environments, mark attendance automatically and provide lecturers with a clear record of every session
Real-time AI
Face detection and recognition using SCRFD and ArcFace
Higher accuracy
Reduced proxy attendance through biometric verification
Lecturer friendly
Attendance monitoring designed for fast class-level review
AI Tuned
GAN AI model fine tunes the detector threshold to improve accuracy
Objectives
Designed with classroom reality in mind
AI-powered attendance
Develop an attendance system for practical classroom environments
Face detection pipeline
Implement real-time face detection and recognition with SCRFD and ArcFace
Proxy prevention
Improve attendance accuracy by verifying identity through facial recognition
Dashboard concept
Provide lecturers with a clean interface for reviewing attendance records
Key Features
Built for speed, trust, and clarity
Real-time face recognition
Detects and identifies students during attendance capture
Automated attendance logging
Records attendance without repeated manual input from lecturers
Fraud prevention
Helps reduce proxy attendance through identity-based verification
Attendance dashboard
Summarizes present, absent, pending, and verified student records
Timestamped records
Stores when attendance was captured for session review and reporting
Smart classroom integration
Designed for future use with cameras, campus systems and access control
GAN AI generator
GAN AI model generates synthetic face images to fine-tune the recognition model
How It Works
A simple flow from face capture to attendance record
1
Camera captures face
The system receives a classroom image or video frame
2
SCRFD detects and aligns
The face is located, cropped and prepared for recognition
3
ArcFace creates embeddings
The face is converted into a compact identity representation
4
Database comparison
The embedding is matched against stored student profiles
5
Attendance is marked
A verified attendance record is created automatically
Technology Stack
Practical AI and web tools for a complete prototype
Python
OpenCV
StyleGAN
SCRFD
ArcFace
Flask / FastAPI
PostgreSQL / MongoDB
HTML
CSS
JavaScript
Market Relevance
Aligned with real-world identity and access systems.
Facial recognition is already used in smart campuses, security systems, access control, e-gates and identity verification. This project brings that direction into a focused academic use case
Classrooms
Lecture halls
Labs
Smart campus access
Corporate training
Team / Contact
Ready for deployment and real-world use
Our solution is designed for real-world deployment, supporting organizations that are looking to modernize their attendance systems with AI-driven automation
Project Lead: Yu ZhangHao
AI Engineer: Dominic Whye / Zhang ChengWei
Computer Vision & Backend Developer: Zhang JiQian
QA Engineer & Documentation Lead: Zhao ShiYin
Request Project Demo