Study on Iris Segmentation Method Using Chan-Vese Model

Abstract:

The first and critical step in the process of an iris recognition system is iris segmentation. Firstly, we detailedly describe the process of pupil and iris localization based on Chan-Vese model. Secondly, we describe the process of unwrapping iris annule region, and obtain rectangular image with the same width but different height. Thirdly, cut rectangular iris image to get normalized image. Fourthly, Multi-channel 2D Gabor, 1-D wavelets and zero-crossing methods were used to extract feature; consequently, decidability indexes of intra-class and inter-class were obtained. Finally, comparatively analyze the pros and cons of the proposed method. Three public iris images databases were taken as experimental samples, the experimental results on these image samples demonstrate that the proposed algorithm has certain advantage.

• Periodical: Advanced Materials Research (Volumes 753-755)
• Pages: 2995-2999
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.753-755.2995
• Online since: August 2013
• Authors: Ying Chen, Feng Yu Yang
• Keywords: Decidability Index, Iris Segmentation, Level Set

References

[1] J. Daugman: Int J Comput Vision, Vol. 45 (2001), pp.25-38.

[2] T.N. Tan, Z.F. He, Z.Z. Sun: Image Vision Comput, Vol. 28 (2010), pp.223-230.

[3] N.B. Puhan, N. Sunha, A.S. Kaushalram: Signal Image Video P, Vol. 5 (2011), pp.105-119.

[4] S. Shah, A. Ross: IEEE T Inf Foren Sec, Vol. 4 (2009), pp.824-836.

[5] K. Roy, P. Bhattacharya, C.Y. Suen: Opt Laser Eng, Vol. 49 (2011), pp.578-588.

[6] T.F. Chan, L.A. Vese: IEEE T Image Process, Vol. 10 (2001), pp.266-277.

[7] T. Dietenbeck, M. Alessandrini, D. Friboulet: In IEEE International Conference On Image Processing, Hong Kong, China, (2010).

[8] L. Ma, T.N. Tan, Y. H Wang: IEEE T Image Process, Vol. 13 (2004), pp.739-750.

[9] W.W. Boles: Eng Appl ARTIF INTEL, Vol. 11 (1998), pp.77-85.

[10] M. Nabti, A. Bouridane: Pattern Recogn, Vol. 41 (2008), pp.868-879.