AN ALGORITHM FOR IRIS IDENTIFICATION USING FOURIER DESCRIPTORS

AN ALGORITHM FOR IRIS IDENTIFICATION USING FOURIER DESCRIPTORS. IEEETEM2003.

[img]
Preview
PDF
s7p3.pdf

Download (97kB) | Preview

Abstract

In this paper, a new biometric identification approach based on the human iris is proposed. The main idea of the technique is to represent the features of the iris with the extracted contour of the inner iris. The extracted boundary includes a wealth of information about the person to allow accurate identification. The technique is fundamentally different from previous techniques that assume the circulatory nature of the inner iris. The iris extracted contours are normalized, the Fourier descriptors are computed, a feature vector is extracted and fed to a statistical classifier. Preliminary results showed a classification success of around 96%. The technique is robust to additive noise, cropping, scale, rotation, translation, and is computationally efficient.

Item Type: Article
Department: College of Engineering and Physics > Electrical Engineering
Depositing User: Users 4447 not found.
Date Deposited: 02 Jun 2008 08:10
Last Modified: 01 Nov 2019 13:27
URI: https://eprints.kfupm.edu.sa/id/eprint/1481