| Title |
Attendance Verification System by using Anti-spoofing Method based on Deep Learning and Nudge Theory |
| Authors |
라이언(I Eon Na) ; 여동훈(Dong Hun Yeo) ; 황병일(Byeong Il Hwang) ; 김동주(Dong Ju Kim) ; 서영주(Young Joo Suh) ; 황도경(Do Kyung Hwang) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.11.78 |
| Keywords |
Anti-spoofing; Face identification; Deep learning; Nudge effect; Eye tracking |
| Abstract |
A novel, practical and universal method of face spoofing detection has been proposed to prevent face spoofing problems in the face identification attendance system by using deep learning and Nudge theory. The proposed method detects face spoofing event while the attendance system is running and figure out whether a person’s purpose has a natural drive to get face identification or not by nudge theory and the new iris detection algorithm to detect and prevent face spoofing situations. Prior face identification attendance system does not have face anti-spoofing technology in their approach. Most systems use only deep learning to judge the situation by extracting features from the nearby network, even if they have. So the plan is feeble on blur situation or has poor performance with an insufficient dataset for network train. Because of this weakness in anti-spoofing, it can not adequately supply AI-based attendance. The proposed method uses Nudge theory and deep learning to upgrade the prior attendance system with a new practical way of a person’s purpose judge system. It uses datasets that have reliability for deep learning network train to have enough performance on face spoofing detection. Our system evaluated various objectivity cases to get reliability and finally accelerated the popularity of AI-based face identification attendance systems for our society. |