Investments in Deep Learning Techniques for Improving the Biometric System Accuracy

Recently, user identification is an essential foundation for protecting information in several applications. However, the need for heightened this identification has expanded the research to focus on the biometric traits of the users. Traditionally, identity determination is assured by recognizing personals used classical security means such password and/or card. Recently, a natural and reliable solution to this problem is offered by biometrics technologies. So, among several biometric modalities, these extracted from the hand have been systematically used to make identification for last years. In other hand, all issues related to the final conception of a biometric system are generally related to the classification task (classifier used). In this paper, we present a Restricted Boltzmann Machine (RBM) for palmprint (PLP) and palm-vein (PLV) identification which are able to classify precisely these modalities. In addition, in order to improving the identification system accuracy, a RBM based Deep Belief Nets (DBN) is also presented. The obtained results, using databases of 400 persons, have showed that deep learning methods has higher performances compared to the classical methods developed in the literature in terms of systems accuracies.

[1]  Geoffrey E. Hinton,et al.  Restricted Boltzmann machines for collaborative filtering , 2007, ICML '07.

[2]  Geoffrey E. Hinton,et al.  Application of Deep Belief Networks for Natural Language Understanding , 2014, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[3]  Sung Bum Pan,et al.  Improvement of Fingerprint Verification by Using the Similarity Distribution , 2013, ITCS.

[4]  Nicolas Le Roux,et al.  Representational Power of Restricted Boltzmann Machines and Deep Belief Networks , 2008, Neural Computation.

[5]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.

[6]  L. Wang,et al.  Some issues of biometrics: technology intelligence, progress and challenges , 2012, Int. J. Inf. Technol. Manag..

[7]  Arun Ross,et al.  Handbook of Multibiometrics , 2006, The Kluwer international series on biometrics.

[8]  Christian Igel,et al.  Training restricted Boltzmann machines: An introduction , 2014, Pattern Recognit..

[9]  Jürgen Schmidhuber,et al.  Evolving deep unsupervised convolutional networks for vision-based reinforcement learning , 2014, GECCO.

[10]  Jürgen Schmidhuber,et al.  Deep learning in neural networks: An overview , 2014, Neural Networks.

[11]  S. Kanade,et al.  Multi-biometrics based cryptographic key regeneration scheme , 2009, 2009 IEEE 3rd International Conference on Biometrics: Theory, Applications, and Systems.

[12]  Jinfeng Yang,et al.  Personal identification based on finger-vein features , 2011, Comput. Hum. Behav..

[13]  David Zhang,et al.  Palmprint feature extraction using 2-D Gabor filters , 2003, Pattern Recognit..

[14]  Zhenhua Guo,et al.  An Online System of Multispectral Palmprint Verification , 2010, IEEE Transactions on Instrumentation and Measurement.

[15]  Geoffrey E. Hinton,et al.  Learning and relearning in Boltzmann machines , 1986 .

[16]  Abdallah Meraoumia,et al.  Combining palmprint & Finger-Knuckle-Print for user identification , 2016, 2016 International Conference on Information Technology for Organizations Development (IT4OD).

[17]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..