Previous Works About Iris Recognition Stages
A Brief Survey
Sasan Harifi
Faculty of Electrical, Computer and IT Engineering,
Islamic Azad University,
Qazvin Branch, Iran
sasan.harifi@gmail.com
Azam Bastanfard
Faculty of Media Engineering,
Islamic Republic of Iran Broadcast University,
Tehran, Iran
bastanfard@iribu.ac.ir
Abstract—By increasing security requirements, biometric as a
perfect solution for people identification is used. One of the
biometrics which is considered as the most accurate and reliable
method, is iris biometric. The iris biometric is about analysis of
patterns in the iris texture. The operation can be done in several
stages. Image acquisition, preprocessing and main processing
including segmentation, normalization, feature extraction and
matching are different stages of iris recognition. In this paper the
previous works for different stages are proposed separately and
classified, with the appropriate tables. Since there are many
different methods for iris recognition, this survey covers some of
the significant methods.
Keywords— Iris recognition, Identification, Image acquisition,
Iris challenges, Preprocessing.
I. INTRODUCTION
The human iris has unusually complex structure which has a
lot of information in its context for iris biometric. Today, Due to
the advancement of algorithms, iris recognition has developed.
In general, iris recognition includes image acquisition stage,
preprocessing stage and the main processing stage which the
main processing stages are: segmentation, normalization,
feature extraction and matching. Figure 1 shows a diagram of
iris recognition stages.
Fig. 1. Diagram of iris recognition stages
This paper is about different methods of iris recognition
stages. Rest of the paper is structured as follows: In section II
image acquisition methods and available devices are mentioned.
There is also information about existing databases. In the section
III the preprocessing methods are presented. Section IV
proposed methods in main processing stages of iris recognition,
separately. Section V is brief evaluation, section VI described
challenges in iris recognition and Section VII is conclusion.
II. IMAGE ACQUISITION
A. Approaches to Image Acquisition
Researchers have studied about image acquisition, such as
the wavelength of light, type of light source, the light reflected
from the surface of the iris which is related to imaging sensor,
the lens features, signal to noise ratio, eye safety and image
quality. Current iris recognition systems operate with
illumination in the 720–900nm range. This choice is driven by
the absorption spectrum of melanin and the absorption spectrum
of silicon. Melanin is the organic pigment that predominates in
the coloring of human irises. It is strongly absorbing in the
visible, but much less so in the near infrared (NIR) 800–900nm.
People with dark eyes have more melanin than those with light
eyes. Dark and light irises are more similar in their light
reflecting and absorbing properties in the NIR than they are in
the visible because the absorption of the melanin drops off in the
NIR. Hence, the structural details of a dark iris show much better
contrast in the NIR than in the visible [1].
There are still important issues in research on iris image
acquisition. One of the issues is taking image of iris which is "at
distance" or "at move". Sarnoff demonstrated its Iris on the
Move
TM
portal system at the Biometrics Consortium
Conference. The system was built for the US Government and
was described in detail by Matey [1]. It was capable of acquiring
iris images at distances of approximately 2–3m with the subject
walking at a normal pace. Several copies of the system were
constructed for the US Government. The long distance record
for a system with co-located sensor and illumination was held,
at least for a time, by another system that Sarnoff built for the
US Government. This iris at a distance system is based on a
Meade LX200-RF/10 8 inch reflecting telescope and a long
distance illuminator consisting of an 850 nm LED focused on
target to produce irradiance of approximately 1 mW/cm
2
. There
has not been a publication describing the details of this device,
Iris Recognition
Main Processing
Image Acquisition
Preprocessing
Iris Segmentation
Normalization
Feature Extraction
Matching
Recognition Results
Proceedings of the Fourth International Conference on e-Technologies and Networks for Development, Lodz, Poland, 2015
ISBN: 978-1-4799-8450-3 ©2015 IEEE 6