ISSN

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 22

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

IRIS Signa l Processing and Classification Using Compound Gabor Filter
Prof. P.V. Rama Raju and Mr. CH. Ganesh
Dept. of Electronics & Communication Engineering,
SRKR Engineering College , Bhimavaram – 534 204, A.P.
(Affiliated to Andhra University, A.P., India)
[email protected] , cell: 9010144688 ,g[email protected] ,cell: 9491761366

Abstract – A resourceful iris processing method is suggested . In the sugge sted method the iris segmentation, normalization,
feature extractions are based on the surveillance that the pupil has lower intensity than the iris, and the iris has lower
intensity than the sclera. By noticing the edge between the pupil and the iris and the edge between the iris and the sclera,
the iris area can be parted from pupil and sclera. Then the normalization technique is applied to reduce the dimensional
inconsistencies later encoding technique is applied to extract the unique feature of iris ima ge, and the resulting one –
dimensional (1 -D) signals are quantized using phase quantiz ation . The phase quantization pro cess produces a number of
bits of information at various resolution levels over pixels on the normalized iris. The Hamming distance is em ployed for
classification , if the hamming distance between two templates is less than the set threshold va lue then two templates are
declared to match or else they are not matched. A case identification program does the unique signal analy sis and
completes the pattern classification. This program gives out its output in graphical form which indicates individual
identification and gives a statement which makes an individual case identification .

Keywords – Segmentation, Normalization, Feature extraction, Phase quantization .
1 INTRODUCTION
A dictionary defines image as a “reproduction or
representation of the form of a person or thing .
Digital image can be defined to be a numerical
representati on of an object to be sampled in an equally
spaced rectangular gri d pattern, and quantized in equal
intervals of gray -level as function of two dimensions .
The word is c arrying out for the simpler access
controls to personal authentication systems and it looks
like biometrics may be the answer. Hence human body
can be use d to uniquely identify. Furthermore,
biometrics is best defined as measurable physiological
and/or behavio ral characteristics that can be utilized to
verify the identity of an individual. They include the
following:
? Iris scanning
? Facial recognition
? Fing erprint verification
? Hand geometry
? Retinal scanning
? Signature verification
? Voice verification
1.1 Iris Introduction
The iris (irises ) is a thin, colored ring , circular
structure in the eye, responsible for controlling the
diameter and size of the pupi ls. The iris is called the
“Living password” because of its unique, random
features. It’s always with human being and can’t be
stolen or faked. As such it makes an excellent
biometrics identifier and classifier.

Figure 1 : Horizontal and v ertical view of Iris .
1.2 Eye Pupil
The pupil is a hole located in the center of the iris
of the eye that allows light to enter the retina . It
appears black because most of the light entering the
pupil is absorbed by the tissues inside the eye . In
response to the amount of light entering the eye,
muscles attached to the iris expand or contract the
aperture at the center of the iris, known as the pupil.
The larger the pupil, the more light can enter.

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 23

1.3 Iris Rec ognition methodology

Figure 2: Flow diagram of the iris recognition steps .
To e xamine the recognition system, eye image
will be used as input . Iris recognition relies on the
unique patterns of the human iris to identify or verify
the identity of an indiv idual . Eye localization and iris
segmentation are sounds the same, These processes are
done by using Canny edge detection technique, circular
Hough transform (CHT) a standard computer vision
algorithm, is commonly employed to deduce the radius
and center coordinate s of the pupil and iris regions .The
iris and pupil boundaries can be approximated as
concen tric circles and t he outer boundaries of iris are
detected with the help of center of pupi l. For a
particular circle the change in intensity between
normal pointing toward center and away from center is
measured. The radius having highest change in
intensity is considered as outer boundary.
Once the iris region is successfully segmented
from an eye image, the next stage is to transform the
iris region so tha t it has fixed dimensions in order to
allow comparisons. Iris normalization is done in order
to make the image independent of the dimensions of
the input image. After iris is normalized , the algorithm
is used to encode the iris data. This process extracts
features from the normalized iris images and encodes it
to generate iris templates . The resultant graphs shown
in section4.
For obtaining test parameter, Hamming distance
gives a measure of how many bits are t he same
between two bit templates . Using the Ha mming
distance of two bit patterns, a decision can be made as
to whether the two patterns were generated from
different irises or from the same one.
2 CASE STUD Y AND RESULT ANALYSIS
After processing and analyzing the iris signal,
The template of the each database signal have been
plotted in Matlab as shown in corresponding figure s.
The results are plott ed in terms of scales (Hamming
distance) on the y -axis and the number of images in the
database (sample images ) on x -axis. The matched
result at that point is represented by a blue color bar
with the threshold value . Red bar indicates that
unmatched condition. In this exposition two cases are
considere d under test.
For every Image in the dataset the following values
have been analyzed:
1. Iris radius and its center coordinates .
2. Filter parameters .
3. Templates of irises .
4. Assignment in Databank.
5. Iris signal classification and identification.
3 INT ER CLASS SAMPLE DATABASE
CLASSIFICATION
Case1: Test signal – S1036 L0 7.jpg (Left Eye)
This signal has been taken from th e CASIA iris
image database collected by Institute of Automation,
Chinese Academy of Sciences .

Figure 3 : Test signal left eye template.
After processing and analyzing the iris signal, the
template of the database signal has been plotted as
shown in ab ove figure. In figure 3 the image is being
tested S1036L07 .jpg, name reveals that seventh left
eye sample image of a 36th perso n in the database.

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 24

Figure 4: Matched eye template from the database .
Figure 4 show that the resultant matched template from
the data base . The Hamming distance distribution and
matching condition between the templates are shown
below. The result is plott ed in terms of scales
(Hamming distance) on the y -axis and the number of
images in the database (sample images ) on x -axis.

Fig ure 5 : Tested image template matched with first template in
the database
In figure5 , the blue bar at position25 indicates that
matching condition below the threshold value. Since
database contains the different person’s data, tested
image is matched with t he best compatible image. The
image is being used for test is matched with processed
left eye images in the database corresponds to the same
person, t he Hamming distance between tested template
and matched template is taken as the optimum
threshold value f or the same category. In the above
example the optimum threshold value is set
programmatically about 0.4, it will be differed for each
individual test.
Case2: Test signal – S1094R03 .jpg (Right Eye)
In this case the signal has been taken unlike t he above
case, the right as tested eye.

Figure 6: Test right sample template.
After processing and analyzing the iris signal, the
template of the database signal has been plotted as
shown in above figure. In figure 6 the image is being
tested S1094R03 .jpg, name reveals that third right eye
sample image of a 94th person in the database.
Figure 7 show that the resultant Hamming distance
distribution and not -match condition between the
templates are shown below, all the bar lines exceeds
the threshold value line 0.4 , hence there is no match at
any case, t he result is plotted in terms of scales
(Hamming distance) on the y -axis and the number of
images in the database (sample images) on x -axis.

Figure 7 : Tested image template unmatched with the any of
template s in th e database
The image is being used for test is a right eye not –
matched with any of processed right eye images in the
database corresponds to the same person, which ca n be
observed at bars on x -axis . In the above example the
optimum threshold value is set p rogrammatically about

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 25

0.4 which is shown as horizontal black line, it will be
differed for each individual test. Hence test of above
signal is done successfully and t he corresponding
observations will be discussed in section 4 .
4 EXPERIMENTAL RESULT ANALYS IS
4.1 Filter parameters , Decidability, Template size,
Number of shifts, Number of filters
The best filter parameters were found through
experimentation with the CASIA data set to compare
perform ance. The decidability value determine s the
optimum parameter s, those are wavelength, template
size, number of shifts, number filters.
Where d’ is decidability, mean of the intra -class
distribution ?s and the mean of the inter -class
distribution ?d and also the standard deviation of the
intra -class and inter -class distributions ? s2, and ? 2d
respectively. The results of experiment, various filter
parameters to encode iris templates are observed bel ow
table1 .
Table1: Different filter parameters with one filter with bandwidth 0.5
N

1 15 0.2 642 0.039 0.4734 0.016 6.067
1 16 0.2396 0.039 0.4725 0.012 6.124
1 17 0.3775 0.039 0.4715 0.012 6.173
1 18 0.2575 0.039 0.4705 0.013 6.216
1 19 0.2601 0.039 0.4694 0.014 6.174
1 20 0.2810 0.039 0.4683 0.015 6.130
1 21 0.2950 0.040 0.4671 0.016 6.066

Figure 8: Decidability Vs center wavelength using filter
bandwidth 0.5
One factor, which will significantly influence the
recognition rate, is iris template size. For the CASIA
data set, the optimum template size was found to be
20×240 pixels.
Figure 8 show s that there exits an optimum center
wavelength 18 for the data set, which produces
maximum decidability d ‘ = 6.2160 . Furthermore,
experiment results show that encoding templates with
multiple filters does not produce better decidability
values, ther efore the optimum number of filters is one
and it produces a compact iris template. The optimum
?/f value was found to be 0.50, Higher the decidability,
greater the separation of same class and inter -class
distributions, which allows for more accurate
reco gnition.

Figure 9 : Inter -class comparison with eight shifts
As the number of shifts increases, the mean of the
inter -class distribution will converge to an optimum
value below the statistical value 0.5. It can be seen in
figure 9 . ‘
22 ()
2
sd
sd
d ??
??
? ?
? min? s? s? d? d? ‘d

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 26

Figure 10 : Mean of in ter-class HD distribution Vs Number of
shifts of two images of same class.
With reference to Figure 10 , the CASIA data set
requires 8 shifts to reach its minimum optimum
threshold. However, it is noted that 4 -to-8 shifts are
enough to be able to account for most of the rotational
inconsistencies in the CASIA database.
5 OBSERVATIONS FROM THE PLOTS
1. The template generated for t he initial scales is
multiplication of number of filters used to the
signal processing, order of pixel array .
2. The 3D plot for 20 -by-24 0, color shift indicates
shift of bits from 0 to 1 or 1 to 0 .
3. The decreasing scale number to 10 -by-1000 shown,
easy to compare the given signal for each
processing value.
4. The bar graph is generated with matching metric
looks better and suitable for finding optimum
signal classification .
5. Blue bar indicates the matching condition, red bar
indicates unmatched condition, and the circle line is
the optimum threshold value on y -axis .
6. Inste ad of numerical observations comparison of
signal analysis using the Hammin g distance bar
graph yields better and accurate Identification .
6 CONCLUSION
This exposition has presented an Inter -class iris
recognition sy stem, which was tested using set of
images as databases of greyscale eye images from
different . Firstly, an effic ient iris processing method is
processed. In the processed method the iris
segmentation, normalization, feature extractions are
being done. Then the encoding technique is applied to
extract the unique feature of iris image, and the
resulting one -dimensiona l (1 -D) signals are phase
quantized using different dissimilarity functions. The
phase quantization process produces a bitwise template
containing a number of bits of information at various
resolution levels. The Hamming distance is employed
for classifica tion of iris templates . The Matlab program
gives out its output in graphical form which indicates
individual identification and gives a statement which
makes an individual case identification.
7 FUTURE SCOPE
The system presented in this publi cation was able
to perform Inter -class classification. However there are
still a number of issues which need to be addressed.
First of all, Inter -class classification system to be
implemented accurately. The automatic segmentation
was not perfect, since it could not successfully segment
the iris regions for all of the eye images in the
database. Another extension to the system would be to
interface it to an iris acquisition camera.
REFERENCES
1 http://www.csse.uwa.edu.au/~masekl01/ 2 Recognition of Human Iris Patterns fo r Biometric Identification , Libor Masek The University of Western Australia, 2003 3 http://en.wikipedia.org/wiki/Iris_(anatomy) 4 Canny, John, “A Computational Approach to Edge Detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PA MI-8, No. 6, 1986, pp. 679 – 698. 5 “A Human Identification Technique Using Images of the Iris and Wavelet Transform” W. W. Boles and B. Boashash 6 “Iris Recognition: An Emerging Biometric Technology” RICHARD P. WILDES, MEMBER, IEEE 7 Optimized Daugman’s Algorit hm for Iris Localization Dr. Mohamed A. Hebaishy National Authority for Remote Sensing and Space Science Gozif Titp St., Elnozha Elgididah. Egypt (11769), Cairo. 8 “Recognition of Human Iris patterns for biometric identification”, Libor Masek, Dr.Peter Koves i, The University of western Australis. 9 Exploit modeling for categorization of electrocardiogram signals through wavelet transform. prof.P.V.Ramaraju, V.Malleswara Rao, Mr CH. Ganesh. 6 th international multi conference on intelligent system, IISN -2012, Mar ch 16-18,NIT Klawad,Haryana,India. 10 ECG signal processing and classification via heart rate fortitude in the vicinity of continuous wavelet transforms.prof.P.V.Rama raju, V.Malleswara ra o and Mr CH. Ganesh, IEEE conference, International journal ACNCN -2012,A U college of engineering,visakhapatna.

ISSN(Online):2533 -8945 VOLUME 4 ISSUE 6

PAPER AVAILABLE ON WWW.IJECEC.COM -VOLUME4 -ISSUE6 -NOV -2018 27

Prof. P.V. Rama Raju received his
Masters Degree in Microwave and
Optical Engineering at the M.K.
University, Tamil Nadu, India. He is
a Professor at the Department of
Electronics and Communication
Engineering S.R.K .R. Engineering
College, AP, India. His research
interests include bio medical -signal
Processing, signal processing, VLSI design and microwave
anechoic chambers design. He is author of several research studies
published in national and international journals and conference
proceedings.

Mr CH. Ganesh pursuing his Masters Degree in Communication
Systems at S.R.K.R Engineering college , Bhimavaram, WG,
Andhrapradesh , India. He is a studen t at the Department of
Electronics and Communication Engineering S.R.K.R. Engineering
College, AP, India.