Biometric-Iris Random Key Generator Using Generalized Regression Neural Networks
In this work, we present a new approach to generate cryptographic keys from iris biometric. The main challenge of the general research is to find a suitable method to generate a cryptographic-iris-key every time the same iris information is analyzed, and this key should be different to the key generated for other users. Some problems to reach this goal are the imperfections that occurs in the biometric acquisition process, the features extraction selection and the matching algorithms. In our work, the key is calculated in four steps. First, the iris is located by use of the integrodifferential operators. Second, a set of features are computed by the use of Gabor filtering. Third, these features are divided in groups, depending on number of bits to be generated. In the final step, we generate a bit for each group of features by using a set of generalized regression neural net classifiers. We develop our experiments using a set of noisy images from the UBIRIS database, and the experimental results are very promising.
KeywordsIndependent Component Analysis Iris Image Back Propagation Neural Network General Regression Neural Network Iris Recognition
Unable to display preview. Download preview PDF.
- 2.Davida, G., Frankel, Y., Matt, B.: On Enabling Secure Applications through Off-line Biometric Identification. In: IEEE Symposium on Security and Privacy, pp. 148–157 (1998)Google Scholar
- 5.Lee, H., Noh, S., Bae, K., park, K., Kim, J.: Invariant Biometric Code Extraction. In: Proceedings of the International Symposium on Intelligent Signal Processing and Communication Systems ISPACS 2004, pp. 181–184 (2004)Google Scholar