This repository contains the source code for the Face Anti-Spoofing System, a real-time attendance authentication application designed and implemented as part of my Bachelor's thesis at Yangzhou University.
The system is built around a lightweight MiniFASNet deep learning model, trained on a custom-developed dataset named SARSpoof, specifically curated with spoofing samples from South Asian individuals to reduce bias and improve accuracy.
Key components include:
- A graphical user interface (GUI) for user registration, login, and logout
- Integrated spoof detection and face recognition modules
- Real-world testing across different lighting, background, distance, and accessory scenarios
- Anti-spoofing protection against print attacks and replay attacks
- MiniFASNet Integration: Lightweight and efficient CNN model for spoof detection
- Real-Time Detection: Webcam-based face authentication and spoofing defense
- Logging: Attendance data saved with timestamps in local log files
- Robust Evaluation: Tested against diverse spoofing attempts (replay, print)
- Inclusive Design: Tailored for the South Asian demographic
-
Python 3.8
-
opencv-python,face_recognition,dlib,torch,torchvision -
GUI:
Tkinter -
Dataset: Custom SARSpoof Dataset (200+ samples with attack variations)
IMPORTANT:
The SARSpoof dataset used in this project is currently unpublished and sensitive.
For security and ethical reasons, this repository has been set to private and the dataset is not included here.
Access to the dataset can be requested on a case-by-case basis for academic purposes only.
This system was developed as part of my undergraduate thesis:
"Design and Implementation of Face Anti-Spoofing System"
Department of Software Engineering, Yangzhou University
Supervised by: Dr. Meiling Fang
- Raoha Bin Mejba
- Email: [email protected]