Orientation Recognition of Iris Pattern

  • Authors:

  • Takeshi Kawasaki;Hironobu Takano;Kiyomi Nakamura

  • Affiliations:

  • Graduate School of Engineering, Toyama Prefectural University, Toyama, Japan 939-0398;Graduate School of Engineering, Toyama Prefectural University, Toyama, Japan 939-0398;Graduate School of Engineering, Toyama Prefectural University, Toyama, Japan 939-0398

  • Venue:
  • Neural Information Processing
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

An iris recognition system that employs a rotation spreading neural network (R-SAN net) has been proposed. R-SAN net is suitable for the orientation recognition of concentric circular patterns such as the iris pattern. The orientation recognition characteristics of R-SAN net for the iris pattern have been investigated. However, it has not yet been understood how the orientation recognition performance of R-SAN net is better than that of other methods employed for orientation recognition.In this study, we evaluated the effectiveness of the orientation recognition performance of R-SAN net by comparing its orientation recognition characteristics with those of the moment (MO) method. The standard deviation of the orientation recognition error in the MO method is larger than that in R-SAN net. For an arbitrary input orientation, R-SAN net showed a fairly good orientation recognition performance as compared to the MO method.