We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Combining Wavelet Transform and LBP Related Features for Fingerprint Liveness Detection.
- Authors
Zhihua Xia; Chengsheng Yuan; Xingming Sun; Decai Sun; Rui Lv
- Abstract
Fingerprint authentication system may verify the identity of the user according to the features of fingerprint. It has been widely used in government departments and national security departments. However, there are some security and privacy problems: Fingerprint authentication system may be easily tricked by fake fingers which are produced through using common artificial materials, such as silicon, wood glue and latex. Thus, designing a fingerprint liveness detection module is necessary for authentication system. In this paper, in order to obtain the optimal set of features and recognize fake fingerprints, a new software-based liveness detection approach using a novel fingerprint parameterization based on wavelet transform and local binary pattern (LBP) related features is applied. The performance of our proposed approach has been evaluated through a comparison with several state-of-the-art techniques for fingerprint liveness detection. In addition, the liveness detection method presented in this paper has an extra advantage over previously studied techniques, since only a fingerprint image is needed to judge whether it is real or fake. Experiments have been carried out by adopting standard databases which are taken from the Liveness Detection Competition 2011 and 2013. Besides that, we have also analyzed the performance of our method for the different combination of wavelet decomposition coefficients during the process of training. Finally, classification accuracy of feature vectors is predicted based on a SVM classifier. Experimental results demonstrate that our method can detect the fingerprint liveness with higher classification accuracy. In addition, this study also confirms that multiresolution analysis is a useful tool for the extraction of texture feature during fingerprint images processing.
- Subjects
BIOMETRIC identification; WAVELET transforms; HUMAN fingerprints; FEATURE extraction; DATA security; BACK propagation
- Publication
IAENG International Journal of Computer Science, 2016, Vol 43, Issue 3, p43
- ISSN
1819-656X
- Publication type
Article