XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
Iris Segmentation using Hough Transform and
Daugman’s integro-differential Operator
Jaideep Singh
CSE(hons. IOT)
Chandigarh University
Mohali,PB,INDIA.
ait.17bcs4566@gmail.com
Er.Rajat Tiwari
ECE
Chandigarh University
Mohali,PB,INDIA.
er.rajattiwari@gmail.com
Abstract—this paper presents the iris segmentation using
hough transform to detect the inner (region between iris and
pupil) as well as the outer boundary (region between iris and
scelra ) of the iris. Iris Segmentation is the main sub-process
of iris recognition as wrongly segmented iris can result in
wrong iris localization, thus a failure for the whole
system.The whole proposed system in this paper is
implemented using web technologies ,which makes it easy to
incorporate this system in an online iris recognition system.
Keywords—segmentation, iris, Hough transform,
convolution.
I. INTRODUCTION
Biometrics refers to the identification of an individual
using the traits that he/she processes. It mainly revolves
around two main categories: physiological (fingerprint,
iris, etc.) or behavioral (signature, voice, etc.).
Physiological methods are mostly preferred because they
doesn’t change over time and cannot be altered by human
intentionally. Out of all physiological identification
method ,iris is mostly adopted
Fig 1. Labelled diagram of human eye [1]
Iris is the circular structure made up of tissues, fibers and
pigments [2]. The iris pattern develops early during the
birth of an individual and doesn’t changes with age. The
uniqueness and rigidity of the iris pattern is not the only
reason for wide adoption of iris recognition as faithful
biometric but it has some additions pros such as
contactless recognition, high accuracy and inability to be
copied.
II. SEGMENTATION METHODS
A. Daugman’s integro differential operator
Daugman in 1993 proposed the first successful iris
segmentation method which uses Integro Differential
operator [3]. This method considers the iris to be a
perfect circle. He applied the Differential operator to
the image domain in search of iris’s outer boundary
first then within the iris’s outer boundary in search of
pupil .The fact that the difference of intensity between
the iris and sclera is more than pupil and iris was
utilized by this method.
Given a pre-processed image, I(x,y), the following
operator can be used to determine the iris’s boundary.
Daugman’s IDO is mathematically expressed as below
(4)
The operator searches over image domain (x,y) for
maximum in the blurred partial derivative with respect
to increasing radius r of the normalized contour integral
of I(x,y) along a circular arc ds of radius r and center
coordinates (xo, yo). The symbol * denotes the
convolution operation and Gσ(r) is a smoothing
function such as a Gaussian of scale σ (standard
deviation) . To normalize the circular integral with
respect to its perimeter, it is divided by 2πr [5]. In
short, this operator behaves as circular edge detector
blurred at a scale set by σ, which searches iteratively
over image space through the parameter set {xo, yo, r}
.First search is for iris’s outer boundary with higher
value of σ . Once the iris’s outer boundary is localized,
the search process with finer value of σ, for the iris
inner boundary is carried out only within the
predetermined region. The computation time associated
with an iris’s outer boundary search process can be
reduced by providing a range of estimates for the
parameter r, that are close to the actual boundary
radius.
B. Circular Hough Transform
Hough transform is a standard image analysis tool for
finding curves that can be defined in a parametrical form