ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 10. Vol. 24. 2018

DOI: 10.17587/it.24.681-688

V. P. Novik, Engineer-Researcher, Iritech Inc., Moscow, e-mail: novikvp@mail.ru, I. A. Matveev, D. Sc., Principal Researcher, e-mail: matveev@ccas.ru, V. Yu. Leonov, PhD, Researcher, Federal Research Center "Computer Science and Control" of Russian Academy of Sciences, Moscow

Matching of Informative Features of Iris Images with the Optimal Path Method

The majority of iris recognition methods use local textural features. The accuracy of iris recognition depends significantly on the precision of local textural features alignment. The article proposes a simple and effective way to enhance the alignment accuracy. Also the problem of choosing the optimal parameters of wavelets forming the iris template is examined. It is shown that improvement of alignment accuracy allows using more informative high-frequency wavelets. Computational experiments were carried out with ICE2005 and CASIA bases from public domain.

Keywords: iris recognition, biometric features, optimal path

P. 681–688

To the contents