Abstract | ||
---|---|---|
Face recognition is one of the most widely used physiological biometrics. However, a printed photo or smartphone recording of a face can easily be presented to a face recognition camera to gain illegitimate access to a system. This paper presents an efficient anti-spoofing approach that can detect whether the face in front of the camera is genuine or fake. The proposed method uses the difference between pairwise discrete cosine transform coefficients and logistic regression as a machine learning algorithm. The experimental results show that the proposed approach outperforms the local binary patterns method, which is a representative technique in this research field. |
Year | Venue | Keywords |
---|---|---|
2015 | 2015 IEEE 5TH INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS - BERLIN (ICCE-BERLIN) | Face anti-spoofing, image block difference, discrete cosine transform, machine learning |
Field | DocType | Citations |
Facial recognition system,Computer vision,Pairwise comparison,Pattern recognition,Computer science,Local binary patterns,Discrete cosine transform,Feature extraction,Artificial intelligence,Biometrics,Logistic regression,Humanoid robot | Conference | 0 |
PageRank | References | Authors |
0.34 | 0 | 1 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hyunho Kang | 1 | 8 | 5.68 |