Palmprint Feature Representation Using Fractal Characteristics.
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Table of Contents
1.
Design, Analysis, and Optimalization of a Microtip Patch Antenna at Frequency 3.55 GHz
for WIMAX Appliaction
Ali El-Alami, Saad Dosse Bennani, Moulhime EL Bekkali, Ali Benbassou ……………….157-162
2. Smart Transisition Power Routing Based on Connectivity and Coverage
in Mobile Adhoc Networks
Bharathiraja.R, S.Karthik, V.P. Arunachalam ….……………………...…………………163-170
3. Design of a Adaptive Distance Learning Hypermedia Based on Learner
Modelling : Application for a Course in Electrical Engineering
Aziz derouich, Karim Mohammed, Driss Marjane, Fayçal Messaoudi ……………..……...171-179
4. Design of Analog Voltage-Mode Multiplier for UHF RFID Passive in
0.18 UM CMOS Process
Smail hassouni, Hassan Qjidaa, Mohamed Latrach .………...……………………………180-195
5. Implementation of Stronger AES by Using Dynamic S-Box Dependent on
Master Key
Sliman Arrag, Abdellatif Hamdoun, Abderrahim Tragha, Salah Eddine Khamlich…………..196-204
6. Efficient High Performance Modified Straight Line Routing for Wireless Sensor
Networks
Dr.R.Kanthavel, R.Dhaya, S.Vimal ………………………...…………………….….205-209
7. Measuring Computer Security Awareness on Internet Banking and Shopping
for Internet Users
Fatimah Sidi, Marzanah A. Jabar, Aida Mustapha, Nor Fazlida Sani, Iskandar Ishak, Siti Rozana
Supian ……………………………………….………………………………….……..210-216
8. Software of Production Scheduling Planning in Manufacture Companies Using Method of
Make to Order
Wiranto Herry Utomo, Aji Lesmana, Hendro Tampake ……………………...…….…....217-222
9. Index Based Steganography: A New Secure Approach for Image Steganography
Using Two Images
B.Persis Urbana Ivy, P.J.Kumar , S.Sureka, G.uma Maheswari ………………..……..….221-228
10. Adaptive Control Using Multiple Models Without Switching
Haisen Ke, Wenrui Li ………………….…………………………………………..….229-235
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
11. Neuro-Genetic Sensorless Sliding Mode Control of A Permanent Magnet
Synchronous Motor Using Leunberger Observer
H. Mahmoudi, A. Essalmi ……………………...…………………………………..….236-246
12. Performance Evaluation of Routing Protocols In Large-Scale Mobile
Ad Hoc Networks
Jamali Abdellah, Naja Najib, El Ouadghiri Driss, Benaini Redouane,
Zyane Abdellah ………………….………………………………………………..….247-254
13. Rule-Based Reasoning Algorithm for Intuitionistic Fuzzy Petri Nets
Arundhati Lenka, DR.Chakradhar Das …………………………………………..….…255-267
14. VM Consolidation Techniques in Cloud Datacenter
T.R.V.Anandharajan, Deepak Bhargavan, M.A. Bhagyaveni. …………...…………...….268-273
15. Palmprint Feature Representation Using Fractal Characteristics
Darma Putra …………...…………………………………………………….….….....274-282
16. Implementation of Virtualization in Data Centers to Increase
Proficiency and Performance
Mueen Uddin, Asadullah Shah, Raed Alsaqour ………………...…………….…....…....283-290
17. Energy-Aware Node Placement in Wireless Sensor Network Using ACO
Rabindra Ku Jena ………………………………………………………...…….......…291-297
18. The Missile Target Extraction in Ultraviolet Images with Noise and
Burst Interferences
Gao Qina, Zhu Yiing, Zou Ping ………………………….…...………………………..298-306
19. Automatic Selection of Filtering Devices in a Distributed Intrusion Prevension
System
Elmehdi Bendriss, Boubker Regragui …………..……………………………………....307-311
20. Efficient Technique for The Classification of Satellite Images Using Fuzzy
Rule Classifier
S.Prabhu,Dr.D.Tensing
…………....……………………………………………..312-323
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
Indexing and Abstracting Information
Thomson Science Citation Index Expanded (Being Monitored)
Thomson Journal Citation Reports/Science Edition (Being Monitored)
Ulrich's Periodicals Directory
DataBase systems and Logic Programming (DBLP)
EBSCO Publishing USA
Directory of Open Access Journals (DOAJ)
Google & Google Scholar Journals
The Index of Information Systems Journals
Information Technology Resources Collection
ZDNet Australia
NLM Catalog (Medi-Informaitcs)
Computing Research and Education Association of Australasia
CiteSeerx
Elsevier SCOPUS
TOC PremierTM
Computer Science Journals
Computers and Applied Sciences Complete
N|W Switzerland
Microsoft Academic Search
SciVerse Hub
J-Gate
INSPEC (The IET)
E-ISSN: 1817-3195
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Editorial Board
EDITOR IN CHIEF
Prof. NIAZ AHMAD
FCE, MOE, Islamabad, PAKISTAN
EDITORIAL BOARD
Dr. CHRISTEL BAIER
University of Malaysia (UKM) 43600 UKM
Faculty of Computer Science, Institute for
BANGI, MALAYSIA.
Theoretical Computer Science, Technical
University Dresden, GERMANY
Dr. KHAIRUDDIN BIN OMAR
Faculty of Information Science and
Dr. YUSUF PISAN
Technology, Universiti Kebangsaan
Department of Software Engineering, Faculty of
Malaysia, 43600 Bangi Selangor Darul-Ehsan,
Information Technology, University of
MALYSIA.
Technology, Sydney, AUSTRALIA
.
Dr. TENGKU MOHD. BIN TENGKU
Dr. YUXIN MAO
SEMBOK
School Of Computer & Information Engineering
Faculty of Information Science and Technology
Zhejiang Gongshang University, CHINA
Universiti Kebangsaan, Malaysia, 43600 Bangi
Selangor Darul-Ehsan, MALYSIA.
Dr. MUHAMMAD SHER
Faculty of Basic and Applied Sciences,
Dr PRABHAT K. MAHANTI
Department of Computer Science, International
Department of Computer Science and Applied
Islamic University, Islamabad. PAKISTAN.
Statistics (CSAS), Hazen Hall Room 311,
University of New Brunswick, Saint John, New
Dr. ZARINA SHUKUR
Brunswick, CANADA.
Computer Science Dept., Fakulti Teknologi dan
Sains Maklumat, University Kebangsaan
Dr. R. PONALAGUSAMY
Malaysia, 43600 Bangi, MALAYSIA.
Department of Mathematics, National Institute of
Technology, Tiruchirappalli, Tamil Nadu,
Dr. NOR AZAN MAT ZIN
INDIA.
Department of Information Science, Faculty of
Information Science & Technology, National
Dr. NITIN UPADHYAY
Computer Science & Information Systems
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
Group, Birla Institute of Technology and Science
E-ISSN: 1817-3195
Electronics, Bani Walid. LIBYA
(BITS), Pilani-Goa Campus, NH-17B Bypass
Road, ZuariNagar, Goa, INDIA.
Dr. A. SERMET ANAGN
Dr. MAUMITA BHATTACHARYA
Eskisehir Osmangazi University, Industrial
SOBIT, Charles Sturt University, Albury - 2640,
Engineering Department, Bademlik Campus,
NSW, AUSTRALIA
26030 Eskisehir, TURKEY.
Dr. SEIFEDINE KADRY
Dr. YACINE LAFIFI
Lebanese International University, LEBONON
Department of Computer Science, University of
Guelma, BP 401, Guelma 24000, ALGERIA.
Dr. AIJUAN DONG
Department of Computer Science, Hood College
Dr. CHRISTOS GRECOS
Frederick, MD 21701. USA
School of Computing, Engineering and Physical
Sciences, University of Central Lancashire,
Dr. S.S.RIAZ AHAMED
Preston PR1 2E, UNITED KINGDOM.
Mohamed Sathak Engineering College,
Kilakarai, & Sathak Institute of Technology,
Dr. JAYANTHI RANJAN
Ramanathapuram , Tamilnadu, INDIA
Institute of Management Technology, Raj Nagar,
Ghaziabad, Uttar Pradesh, INDIA
Dr. ZURIATI AHMAD ZUKARNAIN
University Putra Malaysia, MALAYSIA
Dr. ADEL M. ALIMI
National Engineering School of Sfax (ENIS),
Dr. CHELLALI BENACHAIBA
University of SFAX, TUNISIA
University of Bechar, ALGERIA
Dr. SIKANDAR HAYAT KHIYAL
Dr. MOHD NAZRI ISMAIL
Department of Computer Science, Fatima Jinnah
University of Kuala Lumpur (UniKL)
Women University, Rawalpindi, PAKISTAN
MALYSIA
Dr. ADEL MERABET
Dr. VITUS SAI WA LAM
Department of Electrical & Computer
The University of Hong Kong, CHINA
Engineering, Dalhousie University, Halifax,
CANADA
Dr. WITCHA CHIMPHLEE
Suan Dusit Rajabhat University, Bangkok,
THAILAND
DR. HEMRAJ SAINI
CE&IT Department, Higher Institute of
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Dr. SIDDHIVINAYAK KULKARNI
SHAHBAZ GHAYYUR
University of Ballarat, Ballarat, AUSTRALIA
Faculty of Basic and Applied Sciences,
Department of Computer Science and Software
Dr. S. KARTHIKEYAN
Engineering, International Islamic University,
Caledonian College of Engineering,
Islamabad. PAKISTAN.
OMAN
Dr. T.C.MANJUNATH,
Dr. DRAGAN R. MILIVOJEVIĆ
Professor & Head of the Dept.,Electronicis &
Mining and Metallurgy Institute Bor Zeleni
Communication Engg. Dept,
bulevar 35, 19210 Bor, SERBIA
New Horizon College of Engg.,Bangalore560087, Karnataka, INDIA.
Dr. ABDUL AZIZ
Professor of Computer Science, University of
Dr. Nacer eddine ZAROUR
Central Punjab, PAKISTAN
LIRE Laboratory, Computer Science
Departement, University Mentouri of
Dr.P.DANANJAYAN
Constantine (UMC)
Professor, Department of ECE, PEC, Puducherry,
INDIA.
Dr. RIKTESH SRIVASTAVA
Assistant Professor, Information Systems,
Dr. E. SREENIVASA REDDY
Principal - Vasireddy Venkatadri Institute of
Skyline University College
P O Box 1797, Sharjah, UAE
Technology, Guntur, A.P., INDIA
Dr. Mohd ZAINAL ABIDIN AB
Dr. SANTOSH DHONDOPANT
KHAMITKAR
Ramanand Teerth Marathwada University,
Nanded. Maharashtra 431605, INDIA
Dr. M. IQBAL SARIPAN
(MIEEE, MInstP, Member IAENG, GradBEM)
Dept. of Computer and Communication Systems
Engineering, Faculty of Engineering, Universiti
Putra MALAYSIA
Dr. E. SREENIVASA REDDY
Principal - Vasireddy Venkatadri Institute of
Technology, Guntur, A.P., INDIA
KADIR, PhD, MIEEE
Centre of Excellence on Lightning Protection
(CELP)
Dept. of Electrical and Electronics Engineering,
Faculty of Engineering, UPM,
Selangor,MALAYSIA
Dr. OUSMANE THIARE
Gaston Berger University, Department of
Computer Science, UFR S.A.T
BP 234 Saint-Louis, SENEGAL
Dr. SIDDHIVINAYAK KULKARNI
Graduate School of Information Technology and
Mathematics University of BallartAUSTRALIA
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Dr. BONNY BANERJEE
Dept. of Computer Science & Engineering, Sri
Senior Scientist Audigence, FL, USA, The Ohio
Sai Madhavi Institute of Science & Technology,
State University, Columbus, OH, USA
Mallampudi,
Rajahmundry, A.P, INDIA
Dr. NICKOLAS S. SAPIDIS
Department of Mechanical Engineering,
Dr. KHALID USMANI
University of Western Macedonia
Department of Computer Science, Arid
Kozani GR-50100, GREECE.
Agriculture University, Rawalpindi, PAKISTAN.
Dr. NAZRI BIN MOHD NAWI
Dr. GUFRAN AHAMD ANSARI
Software Engineering Department, Faculty of
Qassim University, College of Computer
Science Computer Information Technology,
Science, Ministry of Higher Education, Qassim
Universiti Tun Hussein Onn
University, KINGDOM OF SAUDI ARABIA
MALAYSIA
Dr. Defa Hu
Dr. JOHN BABALOLA OLADOSU
School of Information, Hunan University of
Ladoke Akintola University of Technology,
Commerce, Changsha 410205, Hunan, P. R. of
Ogbomoso, NIGERIA
CHINA
Dr. ABDELLAH IDRISSI
Department of Computer Science, Faculty of
Science, Mohammed V University - Agdal,
Rabat, MOROCCO
Dr. AMIT CHAUDHRY
University Institute of Engineering and
Technology, Panjab University, Sector-25,
Chandigarh, INDIA
Dr. ASHRAF IMAM
Aligarh Muslim University, Aligarh-INDIA
Dr. MUHAMMAD UMER KHAN
Department of Mechatronics, Faculty of
Engineering, Air University, Islamabad.
PAKISTAN
Dr. MOHAMMED ALI HUSSAIN
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
PALMPRINT FEATURE REPRESENTATION
USING FRACTAL CHARACTERISTICS
DARMA PUTRA
Departement of Electrical Engineering and Information Technology, Faculty of Engineering,
Udayana University, Bali - Indonesia
E-mail: ikgdarmaputra@gmail.com
ABSTRACT
This paper introduced some new techniques to extract the palmprint features based on fractal codes and
fractal dimension. Four new techniques based on fractal, namely FDFC (fractal dimension from fractal
codes), FDIFC (fractal dimension image from fractal code), DIFC (distance image from fractal code), and
AIFC (Angle image from fractal code) are proposed.. The main idea of these methods is the forming
palmprint image representation by using fractal code parameters. There are three parameters of fractal
codes are used namely position of range block, position of domain block, and size of range block. Those
methods have been applied for palmprint verification system. The system has been tested by using database
of 1050 palmprint images, are generated from 5 samples from each of the 210 persons randomly selected.
Experiment results show that our proposed methods can achieve an acceptable accuracy rate.
Keywords: Biometrics, Fractal Codes, Fractal Dimension, Feature Extraction, Palmprint Recognition.
1.
INTRODUCTION
Palmprint is the relatively new in
physiological biometrics. In the palmprint, some
kinds of features could be listed as geometry
features (e.g., width, length, and area of a palm),
palm line features (e.g., principal lines and
wrinkles), and points features (e.g., minutiae and
delta points)[13]. A palm lines have several
advantages compared to other available features:
low-resolution images can be used, low cost
capture devices can be used, it is very difficult or
impossible to fake palmprint, and their
characteristics are stable and unique [9].
The main problem in palmprint
recognition system is how to extract the palmprint
features. Recently, many verification
or
identification technologies using palmprint
biometrics have been developed. Zhang et al. [1],
Darma Putra et al [4][6] applied 2-D Gabor filter to
obtain the texture features of palmprints. Pang at al.
[11] used the pseudo-orthogonal moments to
extract the features of palmprint. LI et al. [12]
transformed the palmprint from spatial to frequency
domain using Fourier transform and then computed
ring and sector energy features. Connie at al.[15]
extracted the texture feature of palmprint using
PCA and ICA. Wu et al.[9] extracted line feature
vectors (LFV) using the magnitudes and
orientations of the gradient of the points on palm-
lines. Kumar et al.[14] combined the palmprints
and hand geometries for verification system. Each
palmprint was divided into overlapping blocks and
the standard deviation value of each block was used
to form the feature vector.
In this paper, we introduce some new
techniques to extract the features of palmprint
based on fractal codes and fractal dimension. The
structures of palm lines are natural and irregular,
while fractal is good tools for the natural and
irregular problems model. That is the reason why
fractal is used to extract the palmprint features.
There are four new methods are proposed namely,
FDFC, FDIFC, DIFC, and AIFC.
All of palm images are used in this paper
are captured using Sony DSC P72 digital camera
with resolution of 640 x 480 pixels. Each person
was requested to put his/her left hand palm down
on with a black background. There are some pegs
on the board to control the hand oriented,
translation, and stretching. ROI of palm images
with 128x128 pixel are extracted automatically by
using two steps in center of mass method that used
in [4][6]. A sample of the hand and pegs position
on the black board and ROI extraction steps are
shown on Figure 1
The rest of this paper is organized as
follows: Section 2 gives a description of fractal
codes and fractal dimension. In section 3, we
detailed our methods to extract the palmprint
274
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
features based on fractal. The palmprint feature
matching is described in section 4. Section 5 reports
our experimental results for both verification and
identification. Finally, the conclusion and future
work are presented in section 6.
(a)
(b)
(c)
(d)
(e)
(f)
E-ISSN: 1817-3195
w i is contracted mapping that describes the
similarity relation between R i and D i , and is usually
defined as an affine transformation as bellow:
xi a i
wi y i = ci
z i 0
bi
di
0
0 x i ei
0 y i + f i
si z i oi
(2)
where x i and y i represent top-left coordinate of the
R i , and z i is the brightness value of its block.
Matrix elements a i , b i , c i , and d i , are the
parameters of spatial rotations and flips of D i , s i is
the contrast scaling and o i is the luminance offset.
Vector elements e i and f i are offset value of space.
We used the size of domain region twice the range
size, so the values of a i , b i , c i , and d i are 0.5. The
actual fractal code f i bellow is usually used in
practice[3].
((
)(
)
)
f i = x Di , y Di , x Ri , y Ri , sizei ,θ i , si , oi (3)
(
Di
)
θ
Figure 1: Extraction of palmprint, (a) original image, (b)
binary image of (a), (c) object bounded, (d) and (e)
position of the first centroid mass in segmented binary
and gray level image, respectively, (f) and (g) position of
the second centroid mass in segmented binary and gray
level image, respectively.
2.
) and (x
, y Di represent top-left
coordinate position of the range block and domain
block, respectively, size is the size of range block,
where x R , y R
i
i
(g)
FRACTAL CHARACTERISTICS
Self-similarity and fractal dimension are
two main characteristics of fractal. These
characteristics become the foundation of the
proposed feature extraction methods in this paper.
This section describes these characteristics.
2.1 Self-Similarity
This characteristic usually is used to
produce the fractal codes of image. Fractal codes of
images are obtained using the partitioned iterated
function system (PIFS) method. PIFS method has
become the foundation of present fractal image
coding. This method works based on local selfsimilarity of image. In PIFS method, each image is
partitioned into its range blocks and domain blocks.
The size of the domain blocks is usually larger than
the size of the range blocks. We applied quadtree
partition technique to divide the image into domain
and range blocks. The relation between a pair of
range block (R i ) and domain block (D i ) is noted as
(1)
R i = wi ( D i )
and i represent the index of the spatial rotation of
the domain region that have values: 0 (no
transformation), 1 (reflection in the y axis), 2
(reflection in the x axis), 3 (180o rotation), 4
(reflection in the line y = x), 5 (rotate 90o counterclockwise), 6 (rotate 90o clockwise), and 7
(reflection in the line y = -x). The fractal codes of a
palmprint image is denoted as follow:
F = fi
N
i =1
(4)
where N represents the number of the fractal code.
The inequality expression bellow is used to indicate
weather the range and the relevant domain block is
similar or not.
d (R, D ) ≤ ε ,
(5)
where d(R,D) represents distance (RMSE) value
between range and domain block, and є is the
tolerance value. The range and the relevant domain
block is similar if d(R,D) is less or equal than є.
Otherwise, the block is regarded not similar.
2.2 Fractal Dimension
275
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
The fractal dimension is computed by
using the box counting method as bellow [8].
(6)
D(s ) = log(N (s )) log(s )
The performance of FDFC is depend on the
selection of tolerance value (є) and M. Figure 2
show the result for each step of FDFC method.
where N(s) represents number of boxes with size s
that containing any pixels object. The number of
N(s) depends on the choice of s. The value of s is
changed from 1 to 2k pixels and counted the
corresponding number N(s), where k = 0, 1,
2,…and 2k < image size. The diagram log(N(s)) (as
y axis) and log(s) (as x axis) is plotted. A straight
line is then fitted to the plotted line in the diagram
and the slope of the line is denoted as dimension
fractal of the object.
3.
E-ISSN: 1817-3195
(a)
PALMPRINT FEATURE
REPRESENTATION
This section presents palmprint feature
extraction based on fractal dimension, and fractal
code to extract the palmprint features. Four new
techniques based on fractal, namely FDFC, FDIFC,
DIFC, and AIFC are proposed.. The main idea of
these methods is the forming palmprint image
representation by using fractal code parameters,
where this is not done in [8].
3.1 FDFC (fractal dimension from fractal codes)
The steps of this method can be explained
as follow.
a. Forming binary image representation
The binary image can be formed as follows.
A( j , k ) = c,
j = 1,2,3, M 1 ,
k = 1,2,3, M 2
c = 1, if j = x and k = y ,
Ri
Ri
otherwise c = 0
3.2 FDIFC (Fractal Dimension Image From
Fractal Codes)
The steps to extract the features of
palmprint by using FDIFC method can be detailed
as follows.
a. Forming binary image representation
This step is same as step (a) in FDFC method.
b. Filtering binary image
Binary image A is filtered by h mxn as follows.
(7)
(8)
(
)
where A represents the binary palmprint image,
M 1 and M 2 represent image width and height (in
this paper, M 1 = M 2 = 128), and x R , y R
i
i
represent top-left coordinate of the range block.
b. Dividing binary image into sub blocks
The binary image A is divided into M x M
blocks, and then fractal dimension value of each
block is computed to form palmprint feature
vector
v = (d 1 , d 2 , d 3 , ..., d M 2)
(b)
(c)
Figure 2: FDFC method. (a) original palmprint image,
(b) binary palmprint image, (c) palmprint feature
representation
A' (x, y ) = A( x, y ) ∗ h( x, y )
(10)
A D ( x, y ) = log( A' ( x, y )) log(s )
(11)
where ∗ represent convolution process, and
h mxn . represents the filter with size m x n which
all its element values are 1 (one).
c. Forming the fractal dimension image
Fractal dimension image is an image which
each of its pixels represents the fractal
dimension value of its neighborhood pixels. The
image can be formed as follows.
(9)
s = max(m, n )
(12)
where AD represent fractal dimension image.
The value of s is assigned to maximum value of
m and n, so the value of fractal dimension is not
greater than two (2).
where d i represent fractal dimension of ith block.
276
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
d. Dividing fractal dimension image into sub
blocks
The fractal dimension image A is divided into
M x M blocks, each of blocks is computed its
intensity mean value, and these value is used to
form the feature vector:
v = µ , µ , µ ,..., µ
1
2
3
2
M
(13)
where µ i represent mean value of block ith.
(a)
E-ISSN: 1817-3195
block. The distance image is not binary image
representation.
b. Filtering the distance images using h mxn
(similar with FDIFC (b))
c. Dividing filtered distance image into sub blocks
The filtered distance image is divided into M x
M blocks, each of block is computed its
intensity mean value, and these value is used to
form the feature vector
vJ = µ , µ , µ ,..., µ
1
2
3
2
M
(16)
where µ i represent mean value of block ith.
DIFC method is depend on the selection of
tolerance value (є), filter size (m and n), and M
parameters. Figure 4 show the result of DIFC
method.
(b)
(a)
(c)
(d)
Figure 3: FDIFC method. (a) original palmprint image,
(b) binary palmprint image, (c) fractal dimension image,
and (d) palmprint feature representation
(b)
FDIFC method is depend on the selection
of tolerance value (є), filter size (m and n), and M
parameters. The result of each step of FDIFC
method is shown in figure 3.
3.3 DIFC (Distance Image From Fractal Code)
The steps of this method can be explained
as follow.
a. Forming distance and direction images
The distance image A can be formed by rules
bellowed.
A( j , k ) = d i ,
di =
if j = x
(
Ri
(
x
j = 1,2,3, M 1 ,
k = 1,2,3, M 2
−x
)2 + (
y
−y
(c)
(d)
Figure 4: DIFC method. (a) original palmprint image,
(b) distance image, (c) filtered image, and (d) palmprint
feature representation
3.4 AIFC (angle image from fractal code)
The method AIFC has been described in
[2]. The first step of this method is the forming of
angle image A as follows
A( j , k ) = α i , j = 1,2,3, M 1 ,
k = 1,2,3, M 2
)2
(14)
and k = y , otherwise d i = 0 . (15)
Ri
where x D , y D
i
i
)
Di
Ri
Di
α i = arctan
Ri
represent top-left coordinate
of the domain block (see formula (8)) and d i
represent distance between range and domain
277
if
j=x
Ri
yD − yR
xD − xR
and k = y
Ri
, otherwise, α i = 0
(17)
i
(18)
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
(
ISSN: 1992-8645
where x D , y D
i
i
)
www.jatit.org
represent top-left coordinate of
the domain block (see formula (8)) and d i represent
the angle between range and domain block. The
angle image is not represents binary image. The
criterion bellow are added to compute the direction
αi.
if
if
if
if
if
if
x R < x D and y R ≥ y D
x R > x D and y R ≥ y D
x R > x D and y R ≤ y D
x R < x D and y R ≤ y D
x R = x D and y R ≥ y D
x R = x D and y R ≤ y D
then α i = α i
then α i = 180 − α i
then α i = 180 + α i
then α i = 360 − α i
then α i = 90
then α i = 270
The rest of the steps of this method are
similar with the step (b) and (c) in DIFC method.
AIFC method is depend on the selection of
tolerance value (є), filter size (m and n), and M
parameters. The result of each step of FDIFC
method is shown in figure 5.
(a)
E-ISSN: 1817-3195
where size represent the size of range blocks (see
formula (8)), and L represent quadtree
decomposition level. The maximum value (K) of
the level L can be computed as follows.
2 K = min( M 1 , M 2 )
in this case K = 7 because M 1 = M 2 = 128.
Figure 2(b), 4(b), and 5(b) show the
palmprint image representation by using the
criterion with L = K and figure 3(b) is the image
without the criterion.
All methods have been explained above
depend on the selection of tolerance value (є).
Figure 6 show the impact of the tolerance value to
the fractal dimension image representation.
According to the figure, the smaller tolerance value
produced more detail information and the higher
tolerance produced less detail information.
The main advantage of our proposed
methods is the palmprint features can be obtained
directly from the fractal codes palmprint image. Its
mean we can form the palmprint feature directly
from the compressed palmprint image because the
compressed image is collection of fractal codes.
(b)
(a)
(c)
(d)
Figure 5: AIFC method. (a) original palmprint image, (b)
angle image, (c) filtered image, and (d) palmprint feature
representation
3.5 Criterion
The binary image, distance image, and
angle image representation are formed based on the
position of range block. The criterion bellow can be
used to filter the range blocks that will be used to
form those images.
sizei = L − 1
(20)
(19)
(b)
(d)
(e)
Figure 6: Fractal dimension image with various
tolerance є, (a),(b),(c), and (d), represent fractal
dimension image with tolerance 2,3,4, and 5,
respectively.
4.
PALMPRINT FEATURE MATCHING
The degree of similarity between two
palmprint features is computed as follows:
278
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
d rs
www.jatit.org
(xr − xr )(xs − xs )T
(21)
= 1−
[(xr − xr )(xr − xr )T ]12 [(xs − xs )(xs − xs )T ]12
5.
where x r , x s are the mean of palmprint feature x r
and x s respectively. The above equation computes
one minus normalized correlation between
palmprint feature vector x r and x s . The value of d rs
is between 0 – 2. The d rs will be closed to 0 if x r
and x s obtained from two image of the same
palmprint. Otherwise, the d rs will be far from 0.
Figure 7 shows three groups of palmprint
from the same palm and palms with
similar/different line structures and their averages
score are listed in Table 1. The matching score of
group A are close to 0, and the matching score of
group B and C are far from 0. It is easy to
distinguish group A from group B and C using
these scores.
Table 1:Matching Score of groups A, B, and C in figure 7
Metode
FDFC
FDIFC
DIFC
AIFC
Group A
0,1873
0,1093
0,1679
0,1762
Group B
0,6029
0,5462
0,6563
0,5057
Group C
0,7121
0,5763
0,8911
0,6425
(a1)
(a2)
(a3)
Group A: palms from the same person
(b1)
(b2)
(b3)
Group B: palms from different person with similar line
structure
E-ISSN: 1817-3195
EXPERIMENT RESULTS
Those methods have been applied for
palmprint verification system. Verification system
is tested using database of 1050 palmprint images,
are generated from 5 samples from each of the 210
persons randomly selected. The averages of the first
three images from each user were used for training
and the rest were used for testing.
The performance of verification system is obtained
by matching each of testing palmprint images with
all of the training palmprint images in the database.
A matching is noted as a correct matching if the
two palmprint images are from the same palm and
as incorrect if otherwise.
Figure 8 (a) show the probability
distributions of genuine and imposter parts of
FDFC method used tolerance = 3, and feature
vector length = 256 (16 x 16 blocks). The genuine
and imposter parts are estimated from correct and
incorrect matching scores, respectively. Figure 8(b)
show its curve characteristics. The FAR, FRR, and
EER of the system are 0.3226%, 2.0734%, and
1.4523% respectively. We also tested the FDFC
method with tolerance=2.5 and the result show the
FAR=0.2312% and FRR=1.4423%.
Figure 9 (a) show the probability
distributions of genuine and imposter parts of
FDIFC method used tolerance = 3, filter size 5 x 5,
and feature vector length = 256 (16 x 16 blocks).
Figure 9(b) show its curve characteristics. The
FAR, FRR, and EER of the system are 0.344%,
0.479%, and 0.4092% respectively.
FAR and FRR values of FDIFC method in
various filter size and feature vector length with
tolerance = 3 are shown in table 2.
Figure 10 (a) show the probability
distributions of genuine and imposter parts of DIFC
method used tolerance = 3, filter size 5 x 3, and
feature vector length = 256 (16 x 16 blocks). Figure
10(b) show its curve characteristics. The FAR,
FRR, and EER of the system are 0.8445%,
1,7544%, and 1.1164% respectively.
Figure 11 (a) show the probability
distributions of genuine and imposter parts of DIFC
method used tolerance = 3, filter size 5 x 3, and
feature vector length = 256 (16 x 16 blocks). Figure
11(b) show its curve characteristics. The FAR,
FRR, and EER of the system are 0,6743%,
1,7544%, and 1.2759% respectively.
(c1)
(c2)
(c3)
Group 3: palms from different person with different line
structure
Figure 7: Groups of palmprint based on typical of
principal lines
279
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
(a)
(b)
Figure 9: (a) Genuine and impostor distributions of
FDIFC (b) its curve characteristics
(b)
Figure 8: (a) Genuine and impostor distributions of
FDFC, (b) its curve characteristics
(a)
(a)
(b)
Figure 10: (a) Genuine and impostor distributions of
DIFC (b) its curve characteristics
280
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Table 2: FAR and FRR of FDIFC method in various filter
size and feature vector length with tolerance = 3
6.
Feature
Vector
Length
(8 x 8)
FRR
FAR
FRR
FAR
FRR
FAR
0.638
0.609
0.479
0.563
0.479
0.463
(16 x 8)
0.638
0.399
0.479
0.574
0.479
0.489
(16 x 16)
0.638
0.529
0.479
0.504
0.479
0.344
(32 x 16)
0.479
0.675
0.479
0.451
0.479
0.304
(32 x 32)
0.638
0.674
0.479
0.675
0.479
0.344
In this paper, we offer new methods
(FDFC, FDIFC, DIFC and AIFC methods) based
on fractal characteristics to represent the palmprint
features. Those methods have been applied for
palmprint verification system. The experiment
results show that the proposed methods can achieve
an acceptable accuracy rate. The FDIFC method
shows the best performance than other proposed
methods.
The main advantage of using those
methods for feature extraction is the methods can
be formed directly by exploitation fractal codes.
Because of fractal is a great method for image
coding and decoding (image compression), so in
the future, we will applied those proposed methods
for palmprint recognition system from compressed
palmprint images.
Filter 3 x 3
Filter 5 x 3
Filter 5 x 5
CONCLUSIONS AND FUTURE WORK
ACKNOWLEDGEMENT
The authors would like to thank
Directorate General of Higher Education of
Indonesia and Udayana University for supporting
this research.
(a)
REFERENCES:
(b)
Figure 11: (a) Genuine and impostor distributions of
AIFC (b) its curve characteristics
FAR and FRR values of DIFC and AIFC
method in various filter size and feature vector
length with tolerance = 3 are shown in table 3.
Table 3: FAR and FRR of DIFC and AIFC method in
various filter size and feature vector length with
tolerance = 3
DIFC
AIFC
Filter Size/
Number of Blocks
FRR
FAR
FRR
FAR
3 x 3 (8 x 8 block)
2,8708
0,7382
2,5518
0,7877
5 x 3 (8 x 8 block)
2,7113
0,8189
2,8708
0,7683
5 x 5 (8 x 8 block)
2,7113
0,7576
2,8708
0,8199
3 x 3 (16x16 block)
1,7544
0,9222
1,7544
0,8925
5 x 3 (16x16 block)
1,7544
0,8445
1,7544
0,6743
5 x 5 (16x16 block)
1,7544
0,8619
1,7544
0,6998
[1] D. Zhang, W.K. Kong, J.You & M.Wong,
Online Palmprint Identification, IEEE
Transaction on Pattern Analysis and Machine
Intelligence, 25(9), 2003.
[2] Darma Putra, PIFS Code based for Biometric
Palmprint Verification, International Journal
of Computer Science and Information Security
(IJCSIS), Vol. 9, No. 2, 2011, p. 47-52.
[3] T. Yokoyama, K. Sugawara & T. Watanabe,
Similarity-based image retrieval system using
partitioned iterated function system codes, The
8th International Symposium on Artificial Life
and Robotics, Oita, Japan, 2006.
[4] Darma Putra, Erdiawan, High Performance
Palmprint Identification System Based On
Two Dimensional Gabor, Telkomnika UAD,
Vol. 8, No. 3, 2010, p. 195-326
[5] A.K. Jain, Ross A. & S. Pankanti, A Prototype
Hand Geometry-based Verification System,
IEEE Trans. On Circuits and System for Video
Technology, 4(1), 1999.
[6] Darma Putra, Wira Bhuana, Erdiawan,
Pembentukan Kode Telapak Tangan (Palm
Code) berbasis Metode Gabor 2D, Makara UI,
Vol. 15 No. 2, 2011, p. 161-167
281
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
[7] D. Zhang, W. Shu, Two novel characteristics
in palmprint verification: datum point
invariance and line feature matching, pattern
recognition, 32(1), 1999, 691-702.
[8] I Ketut Gede Darma Putra, Sistem Verifikasi
Biometrika Telapak Tangan dengan metode
dimensi fraktal dan lacunarity, Teknologi
Elektro UNUD, Vol. 8 No. 2, 2009, p. 1-6
[9] X.Q. WU, K.Q. Wang & D. Zhang, An
Approach to Line Feature Representation and
Matching for Palmprint Recognition, Journal
of Software, 15(6), 2004.
[10] N. Duta, A.K. Jain & K.V. Mardia, Matching
of Palmprints, Pattern Recognition Letters,
23(1), 2002, 477-485.
[11] Y. Pang, T.B.J. Andrew, N.C.L. David & H.F.
San, Palmprint Verification with Moments,
Journal of WSCG, 12(1-3), 2003.
[12] W. LI, D. Zhang & S. XU, Palmprint
Recognition Based on Fourier Transform,
Journal of Software, 13(5), 2002.
[13] W. Shu & D. Zhang, Automated personal
identification by palmprint, Opt. eng., 37(8),
1998, 2359-2363.
[14] A. Kumar, D.C.M.Wong, C.S. Helen, Anil K.
Jain, Personal Verification using Palmprint
and Hand Geometry Biometric, 2004,
http:/biometrics.cse.msu.
edu/Kumar_AVBPA2003.pdf
[16] T. Connie., A. Teoh, M. Goh & D. Ngo,
Palmprint Recognition with PCA and ICA,
2003,
sprg.massye.ac.nz/ivcnz/proccedings/ivcnz_41
.pdf.
282
E-ISSN: 1817-3195
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Table of Contents
1.
Design, Analysis, and Optimalization of a Microtip Patch Antenna at Frequency 3.55 GHz
for WIMAX Appliaction
Ali El-Alami, Saad Dosse Bennani, Moulhime EL Bekkali, Ali Benbassou ……………….157-162
2. Smart Transisition Power Routing Based on Connectivity and Coverage
in Mobile Adhoc Networks
Bharathiraja.R, S.Karthik, V.P. Arunachalam ….……………………...…………………163-170
3. Design of a Adaptive Distance Learning Hypermedia Based on Learner
Modelling : Application for a Course in Electrical Engineering
Aziz derouich, Karim Mohammed, Driss Marjane, Fayçal Messaoudi ……………..……...171-179
4. Design of Analog Voltage-Mode Multiplier for UHF RFID Passive in
0.18 UM CMOS Process
Smail hassouni, Hassan Qjidaa, Mohamed Latrach .………...……………………………180-195
5. Implementation of Stronger AES by Using Dynamic S-Box Dependent on
Master Key
Sliman Arrag, Abdellatif Hamdoun, Abderrahim Tragha, Salah Eddine Khamlich…………..196-204
6. Efficient High Performance Modified Straight Line Routing for Wireless Sensor
Networks
Dr.R.Kanthavel, R.Dhaya, S.Vimal ………………………...…………………….….205-209
7. Measuring Computer Security Awareness on Internet Banking and Shopping
for Internet Users
Fatimah Sidi, Marzanah A. Jabar, Aida Mustapha, Nor Fazlida Sani, Iskandar Ishak, Siti Rozana
Supian ……………………………………….………………………………….……..210-216
8. Software of Production Scheduling Planning in Manufacture Companies Using Method of
Make to Order
Wiranto Herry Utomo, Aji Lesmana, Hendro Tampake ……………………...…….…....217-222
9. Index Based Steganography: A New Secure Approach for Image Steganography
Using Two Images
B.Persis Urbana Ivy, P.J.Kumar , S.Sureka, G.uma Maheswari ………………..……..….221-228
10. Adaptive Control Using Multiple Models Without Switching
Haisen Ke, Wenrui Li ………………….…………………………………………..….229-235
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
11. Neuro-Genetic Sensorless Sliding Mode Control of A Permanent Magnet
Synchronous Motor Using Leunberger Observer
H. Mahmoudi, A. Essalmi ……………………...…………………………………..….236-246
12. Performance Evaluation of Routing Protocols In Large-Scale Mobile
Ad Hoc Networks
Jamali Abdellah, Naja Najib, El Ouadghiri Driss, Benaini Redouane,
Zyane Abdellah ………………….………………………………………………..….247-254
13. Rule-Based Reasoning Algorithm for Intuitionistic Fuzzy Petri Nets
Arundhati Lenka, DR.Chakradhar Das …………………………………………..….…255-267
14. VM Consolidation Techniques in Cloud Datacenter
T.R.V.Anandharajan, Deepak Bhargavan, M.A. Bhagyaveni. …………...…………...….268-273
15. Palmprint Feature Representation Using Fractal Characteristics
Darma Putra …………...…………………………………………………….….….....274-282
16. Implementation of Virtualization in Data Centers to Increase
Proficiency and Performance
Mueen Uddin, Asadullah Shah, Raed Alsaqour ………………...…………….…....…....283-290
17. Energy-Aware Node Placement in Wireless Sensor Network Using ACO
Rabindra Ku Jena ………………………………………………………...…….......…291-297
18. The Missile Target Extraction in Ultraviolet Images with Noise and
Burst Interferences
Gao Qina, Zhu Yiing, Zou Ping ………………………….…...………………………..298-306
19. Automatic Selection of Filtering Devices in a Distributed Intrusion Prevension
System
Elmehdi Bendriss, Boubker Regragui …………..……………………………………....307-311
20. Efficient Technique for The Classification of Satellite Images Using Fuzzy
Rule Classifier
S.Prabhu,Dr.D.Tensing
…………....……………………………………………..312-323
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
Indexing and Abstracting Information
Thomson Science Citation Index Expanded (Being Monitored)
Thomson Journal Citation Reports/Science Edition (Being Monitored)
Ulrich's Periodicals Directory
DataBase systems and Logic Programming (DBLP)
EBSCO Publishing USA
Directory of Open Access Journals (DOAJ)
Google & Google Scholar Journals
The Index of Information Systems Journals
Information Technology Resources Collection
ZDNet Australia
NLM Catalog (Medi-Informaitcs)
Computing Research and Education Association of Australasia
CiteSeerx
Elsevier SCOPUS
TOC PremierTM
Computer Science Journals
Computers and Applied Sciences Complete
N|W Switzerland
Microsoft Academic Search
SciVerse Hub
J-Gate
INSPEC (The IET)
E-ISSN: 1817-3195
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Editorial Board
EDITOR IN CHIEF
Prof. NIAZ AHMAD
FCE, MOE, Islamabad, PAKISTAN
EDITORIAL BOARD
Dr. CHRISTEL BAIER
University of Malaysia (UKM) 43600 UKM
Faculty of Computer Science, Institute for
BANGI, MALAYSIA.
Theoretical Computer Science, Technical
University Dresden, GERMANY
Dr. KHAIRUDDIN BIN OMAR
Faculty of Information Science and
Dr. YUSUF PISAN
Technology, Universiti Kebangsaan
Department of Software Engineering, Faculty of
Malaysia, 43600 Bangi Selangor Darul-Ehsan,
Information Technology, University of
MALYSIA.
Technology, Sydney, AUSTRALIA
.
Dr. TENGKU MOHD. BIN TENGKU
Dr. YUXIN MAO
SEMBOK
School Of Computer & Information Engineering
Faculty of Information Science and Technology
Zhejiang Gongshang University, CHINA
Universiti Kebangsaan, Malaysia, 43600 Bangi
Selangor Darul-Ehsan, MALYSIA.
Dr. MUHAMMAD SHER
Faculty of Basic and Applied Sciences,
Dr PRABHAT K. MAHANTI
Department of Computer Science, International
Department of Computer Science and Applied
Islamic University, Islamabad. PAKISTAN.
Statistics (CSAS), Hazen Hall Room 311,
University of New Brunswick, Saint John, New
Dr. ZARINA SHUKUR
Brunswick, CANADA.
Computer Science Dept., Fakulti Teknologi dan
Sains Maklumat, University Kebangsaan
Dr. R. PONALAGUSAMY
Malaysia, 43600 Bangi, MALAYSIA.
Department of Mathematics, National Institute of
Technology, Tiruchirappalli, Tamil Nadu,
Dr. NOR AZAN MAT ZIN
INDIA.
Department of Information Science, Faculty of
Information Science & Technology, National
Dr. NITIN UPADHYAY
Computer Science & Information Systems
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
Group, Birla Institute of Technology and Science
E-ISSN: 1817-3195
Electronics, Bani Walid. LIBYA
(BITS), Pilani-Goa Campus, NH-17B Bypass
Road, ZuariNagar, Goa, INDIA.
Dr. A. SERMET ANAGN
Dr. MAUMITA BHATTACHARYA
Eskisehir Osmangazi University, Industrial
SOBIT, Charles Sturt University, Albury - 2640,
Engineering Department, Bademlik Campus,
NSW, AUSTRALIA
26030 Eskisehir, TURKEY.
Dr. SEIFEDINE KADRY
Dr. YACINE LAFIFI
Lebanese International University, LEBONON
Department of Computer Science, University of
Guelma, BP 401, Guelma 24000, ALGERIA.
Dr. AIJUAN DONG
Department of Computer Science, Hood College
Dr. CHRISTOS GRECOS
Frederick, MD 21701. USA
School of Computing, Engineering and Physical
Sciences, University of Central Lancashire,
Dr. S.S.RIAZ AHAMED
Preston PR1 2E, UNITED KINGDOM.
Mohamed Sathak Engineering College,
Kilakarai, & Sathak Institute of Technology,
Dr. JAYANTHI RANJAN
Ramanathapuram , Tamilnadu, INDIA
Institute of Management Technology, Raj Nagar,
Ghaziabad, Uttar Pradesh, INDIA
Dr. ZURIATI AHMAD ZUKARNAIN
University Putra Malaysia, MALAYSIA
Dr. ADEL M. ALIMI
National Engineering School of Sfax (ENIS),
Dr. CHELLALI BENACHAIBA
University of SFAX, TUNISIA
University of Bechar, ALGERIA
Dr. SIKANDAR HAYAT KHIYAL
Dr. MOHD NAZRI ISMAIL
Department of Computer Science, Fatima Jinnah
University of Kuala Lumpur (UniKL)
Women University, Rawalpindi, PAKISTAN
MALYSIA
Dr. ADEL MERABET
Dr. VITUS SAI WA LAM
Department of Electrical & Computer
The University of Hong Kong, CHINA
Engineering, Dalhousie University, Halifax,
CANADA
Dr. WITCHA CHIMPHLEE
Suan Dusit Rajabhat University, Bangkok,
THAILAND
DR. HEMRAJ SAINI
CE&IT Department, Higher Institute of
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Dr. SIDDHIVINAYAK KULKARNI
SHAHBAZ GHAYYUR
University of Ballarat, Ballarat, AUSTRALIA
Faculty of Basic and Applied Sciences,
Department of Computer Science and Software
Dr. S. KARTHIKEYAN
Engineering, International Islamic University,
Caledonian College of Engineering,
Islamabad. PAKISTAN.
OMAN
Dr. T.C.MANJUNATH,
Dr. DRAGAN R. MILIVOJEVIĆ
Professor & Head of the Dept.,Electronicis &
Mining and Metallurgy Institute Bor Zeleni
Communication Engg. Dept,
bulevar 35, 19210 Bor, SERBIA
New Horizon College of Engg.,Bangalore560087, Karnataka, INDIA.
Dr. ABDUL AZIZ
Professor of Computer Science, University of
Dr. Nacer eddine ZAROUR
Central Punjab, PAKISTAN
LIRE Laboratory, Computer Science
Departement, University Mentouri of
Dr.P.DANANJAYAN
Constantine (UMC)
Professor, Department of ECE, PEC, Puducherry,
INDIA.
Dr. RIKTESH SRIVASTAVA
Assistant Professor, Information Systems,
Dr. E. SREENIVASA REDDY
Principal - Vasireddy Venkatadri Institute of
Skyline University College
P O Box 1797, Sharjah, UAE
Technology, Guntur, A.P., INDIA
Dr. Mohd ZAINAL ABIDIN AB
Dr. SANTOSH DHONDOPANT
KHAMITKAR
Ramanand Teerth Marathwada University,
Nanded. Maharashtra 431605, INDIA
Dr. M. IQBAL SARIPAN
(MIEEE, MInstP, Member IAENG, GradBEM)
Dept. of Computer and Communication Systems
Engineering, Faculty of Engineering, Universiti
Putra MALAYSIA
Dr. E. SREENIVASA REDDY
Principal - Vasireddy Venkatadri Institute of
Technology, Guntur, A.P., INDIA
KADIR, PhD, MIEEE
Centre of Excellence on Lightning Protection
(CELP)
Dept. of Electrical and Electronics Engineering,
Faculty of Engineering, UPM,
Selangor,MALAYSIA
Dr. OUSMANE THIARE
Gaston Berger University, Department of
Computer Science, UFR S.A.T
BP 234 Saint-Louis, SENEGAL
Dr. SIDDHIVINAYAK KULKARNI
Graduate School of Information Technology and
Mathematics University of BallartAUSTRALIA
Journal of Theoretical and Applied Information Technology
20th July 2013. Vol. 53 No.2
© 2005 - 2012 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Dr. BONNY BANERJEE
Dept. of Computer Science & Engineering, Sri
Senior Scientist Audigence, FL, USA, The Ohio
Sai Madhavi Institute of Science & Technology,
State University, Columbus, OH, USA
Mallampudi,
Rajahmundry, A.P, INDIA
Dr. NICKOLAS S. SAPIDIS
Department of Mechanical Engineering,
Dr. KHALID USMANI
University of Western Macedonia
Department of Computer Science, Arid
Kozani GR-50100, GREECE.
Agriculture University, Rawalpindi, PAKISTAN.
Dr. NAZRI BIN MOHD NAWI
Dr. GUFRAN AHAMD ANSARI
Software Engineering Department, Faculty of
Qassim University, College of Computer
Science Computer Information Technology,
Science, Ministry of Higher Education, Qassim
Universiti Tun Hussein Onn
University, KINGDOM OF SAUDI ARABIA
MALAYSIA
Dr. Defa Hu
Dr. JOHN BABALOLA OLADOSU
School of Information, Hunan University of
Ladoke Akintola University of Technology,
Commerce, Changsha 410205, Hunan, P. R. of
Ogbomoso, NIGERIA
CHINA
Dr. ABDELLAH IDRISSI
Department of Computer Science, Faculty of
Science, Mohammed V University - Agdal,
Rabat, MOROCCO
Dr. AMIT CHAUDHRY
University Institute of Engineering and
Technology, Panjab University, Sector-25,
Chandigarh, INDIA
Dr. ASHRAF IMAM
Aligarh Muslim University, Aligarh-INDIA
Dr. MUHAMMAD UMER KHAN
Department of Mechatronics, Faculty of
Engineering, Air University, Islamabad.
PAKISTAN
Dr. MOHAMMED ALI HUSSAIN
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
PALMPRINT FEATURE REPRESENTATION
USING FRACTAL CHARACTERISTICS
DARMA PUTRA
Departement of Electrical Engineering and Information Technology, Faculty of Engineering,
Udayana University, Bali - Indonesia
E-mail: ikgdarmaputra@gmail.com
ABSTRACT
This paper introduced some new techniques to extract the palmprint features based on fractal codes and
fractal dimension. Four new techniques based on fractal, namely FDFC (fractal dimension from fractal
codes), FDIFC (fractal dimension image from fractal code), DIFC (distance image from fractal code), and
AIFC (Angle image from fractal code) are proposed.. The main idea of these methods is the forming
palmprint image representation by using fractal code parameters. There are three parameters of fractal
codes are used namely position of range block, position of domain block, and size of range block. Those
methods have been applied for palmprint verification system. The system has been tested by using database
of 1050 palmprint images, are generated from 5 samples from each of the 210 persons randomly selected.
Experiment results show that our proposed methods can achieve an acceptable accuracy rate.
Keywords: Biometrics, Fractal Codes, Fractal Dimension, Feature Extraction, Palmprint Recognition.
1.
INTRODUCTION
Palmprint is the relatively new in
physiological biometrics. In the palmprint, some
kinds of features could be listed as geometry
features (e.g., width, length, and area of a palm),
palm line features (e.g., principal lines and
wrinkles), and points features (e.g., minutiae and
delta points)[13]. A palm lines have several
advantages compared to other available features:
low-resolution images can be used, low cost
capture devices can be used, it is very difficult or
impossible to fake palmprint, and their
characteristics are stable and unique [9].
The main problem in palmprint
recognition system is how to extract the palmprint
features. Recently, many verification
or
identification technologies using palmprint
biometrics have been developed. Zhang et al. [1],
Darma Putra et al [4][6] applied 2-D Gabor filter to
obtain the texture features of palmprints. Pang at al.
[11] used the pseudo-orthogonal moments to
extract the features of palmprint. LI et al. [12]
transformed the palmprint from spatial to frequency
domain using Fourier transform and then computed
ring and sector energy features. Connie at al.[15]
extracted the texture feature of palmprint using
PCA and ICA. Wu et al.[9] extracted line feature
vectors (LFV) using the magnitudes and
orientations of the gradient of the points on palm-
lines. Kumar et al.[14] combined the palmprints
and hand geometries for verification system. Each
palmprint was divided into overlapping blocks and
the standard deviation value of each block was used
to form the feature vector.
In this paper, we introduce some new
techniques to extract the features of palmprint
based on fractal codes and fractal dimension. The
structures of palm lines are natural and irregular,
while fractal is good tools for the natural and
irregular problems model. That is the reason why
fractal is used to extract the palmprint features.
There are four new methods are proposed namely,
FDFC, FDIFC, DIFC, and AIFC.
All of palm images are used in this paper
are captured using Sony DSC P72 digital camera
with resolution of 640 x 480 pixels. Each person
was requested to put his/her left hand palm down
on with a black background. There are some pegs
on the board to control the hand oriented,
translation, and stretching. ROI of palm images
with 128x128 pixel are extracted automatically by
using two steps in center of mass method that used
in [4][6]. A sample of the hand and pegs position
on the black board and ROI extraction steps are
shown on Figure 1
The rest of this paper is organized as
follows: Section 2 gives a description of fractal
codes and fractal dimension. In section 3, we
detailed our methods to extract the palmprint
274
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
features based on fractal. The palmprint feature
matching is described in section 4. Section 5 reports
our experimental results for both verification and
identification. Finally, the conclusion and future
work are presented in section 6.
(a)
(b)
(c)
(d)
(e)
(f)
E-ISSN: 1817-3195
w i is contracted mapping that describes the
similarity relation between R i and D i , and is usually
defined as an affine transformation as bellow:
xi a i
wi y i = ci
z i 0
bi
di
0
0 x i ei
0 y i + f i
si z i oi
(2)
where x i and y i represent top-left coordinate of the
R i , and z i is the brightness value of its block.
Matrix elements a i , b i , c i , and d i , are the
parameters of spatial rotations and flips of D i , s i is
the contrast scaling and o i is the luminance offset.
Vector elements e i and f i are offset value of space.
We used the size of domain region twice the range
size, so the values of a i , b i , c i , and d i are 0.5. The
actual fractal code f i bellow is usually used in
practice[3].
((
)(
)
)
f i = x Di , y Di , x Ri , y Ri , sizei ,θ i , si , oi (3)
(
Di
)
θ
Figure 1: Extraction of palmprint, (a) original image, (b)
binary image of (a), (c) object bounded, (d) and (e)
position of the first centroid mass in segmented binary
and gray level image, respectively, (f) and (g) position of
the second centroid mass in segmented binary and gray
level image, respectively.
2.
) and (x
, y Di represent top-left
coordinate position of the range block and domain
block, respectively, size is the size of range block,
where x R , y R
i
i
(g)
FRACTAL CHARACTERISTICS
Self-similarity and fractal dimension are
two main characteristics of fractal. These
characteristics become the foundation of the
proposed feature extraction methods in this paper.
This section describes these characteristics.
2.1 Self-Similarity
This characteristic usually is used to
produce the fractal codes of image. Fractal codes of
images are obtained using the partitioned iterated
function system (PIFS) method. PIFS method has
become the foundation of present fractal image
coding. This method works based on local selfsimilarity of image. In PIFS method, each image is
partitioned into its range blocks and domain blocks.
The size of the domain blocks is usually larger than
the size of the range blocks. We applied quadtree
partition technique to divide the image into domain
and range blocks. The relation between a pair of
range block (R i ) and domain block (D i ) is noted as
(1)
R i = wi ( D i )
and i represent the index of the spatial rotation of
the domain region that have values: 0 (no
transformation), 1 (reflection in the y axis), 2
(reflection in the x axis), 3 (180o rotation), 4
(reflection in the line y = x), 5 (rotate 90o counterclockwise), 6 (rotate 90o clockwise), and 7
(reflection in the line y = -x). The fractal codes of a
palmprint image is denoted as follow:
F = fi
N
i =1
(4)
where N represents the number of the fractal code.
The inequality expression bellow is used to indicate
weather the range and the relevant domain block is
similar or not.
d (R, D ) ≤ ε ,
(5)
where d(R,D) represents distance (RMSE) value
between range and domain block, and є is the
tolerance value. The range and the relevant domain
block is similar if d(R,D) is less or equal than є.
Otherwise, the block is regarded not similar.
2.2 Fractal Dimension
275
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
The fractal dimension is computed by
using the box counting method as bellow [8].
(6)
D(s ) = log(N (s )) log(s )
The performance of FDFC is depend on the
selection of tolerance value (є) and M. Figure 2
show the result for each step of FDFC method.
where N(s) represents number of boxes with size s
that containing any pixels object. The number of
N(s) depends on the choice of s. The value of s is
changed from 1 to 2k pixels and counted the
corresponding number N(s), where k = 0, 1,
2,…and 2k < image size. The diagram log(N(s)) (as
y axis) and log(s) (as x axis) is plotted. A straight
line is then fitted to the plotted line in the diagram
and the slope of the line is denoted as dimension
fractal of the object.
3.
E-ISSN: 1817-3195
(a)
PALMPRINT FEATURE
REPRESENTATION
This section presents palmprint feature
extraction based on fractal dimension, and fractal
code to extract the palmprint features. Four new
techniques based on fractal, namely FDFC, FDIFC,
DIFC, and AIFC are proposed.. The main idea of
these methods is the forming palmprint image
representation by using fractal code parameters,
where this is not done in [8].
3.1 FDFC (fractal dimension from fractal codes)
The steps of this method can be explained
as follow.
a. Forming binary image representation
The binary image can be formed as follows.
A( j , k ) = c,
j = 1,2,3, M 1 ,
k = 1,2,3, M 2
c = 1, if j = x and k = y ,
Ri
Ri
otherwise c = 0
3.2 FDIFC (Fractal Dimension Image From
Fractal Codes)
The steps to extract the features of
palmprint by using FDIFC method can be detailed
as follows.
a. Forming binary image representation
This step is same as step (a) in FDFC method.
b. Filtering binary image
Binary image A is filtered by h mxn as follows.
(7)
(8)
(
)
where A represents the binary palmprint image,
M 1 and M 2 represent image width and height (in
this paper, M 1 = M 2 = 128), and x R , y R
i
i
represent top-left coordinate of the range block.
b. Dividing binary image into sub blocks
The binary image A is divided into M x M
blocks, and then fractal dimension value of each
block is computed to form palmprint feature
vector
v = (d 1 , d 2 , d 3 , ..., d M 2)
(b)
(c)
Figure 2: FDFC method. (a) original palmprint image,
(b) binary palmprint image, (c) palmprint feature
representation
A' (x, y ) = A( x, y ) ∗ h( x, y )
(10)
A D ( x, y ) = log( A' ( x, y )) log(s )
(11)
where ∗ represent convolution process, and
h mxn . represents the filter with size m x n which
all its element values are 1 (one).
c. Forming the fractal dimension image
Fractal dimension image is an image which
each of its pixels represents the fractal
dimension value of its neighborhood pixels. The
image can be formed as follows.
(9)
s = max(m, n )
(12)
where AD represent fractal dimension image.
The value of s is assigned to maximum value of
m and n, so the value of fractal dimension is not
greater than two (2).
where d i represent fractal dimension of ith block.
276
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
d. Dividing fractal dimension image into sub
blocks
The fractal dimension image A is divided into
M x M blocks, each of blocks is computed its
intensity mean value, and these value is used to
form the feature vector:
v = µ , µ , µ ,..., µ
1
2
3
2
M
(13)
where µ i represent mean value of block ith.
(a)
E-ISSN: 1817-3195
block. The distance image is not binary image
representation.
b. Filtering the distance images using h mxn
(similar with FDIFC (b))
c. Dividing filtered distance image into sub blocks
The filtered distance image is divided into M x
M blocks, each of block is computed its
intensity mean value, and these value is used to
form the feature vector
vJ = µ , µ , µ ,..., µ
1
2
3
2
M
(16)
where µ i represent mean value of block ith.
DIFC method is depend on the selection of
tolerance value (є), filter size (m and n), and M
parameters. Figure 4 show the result of DIFC
method.
(b)
(a)
(c)
(d)
Figure 3: FDIFC method. (a) original palmprint image,
(b) binary palmprint image, (c) fractal dimension image,
and (d) palmprint feature representation
(b)
FDIFC method is depend on the selection
of tolerance value (є), filter size (m and n), and M
parameters. The result of each step of FDIFC
method is shown in figure 3.
3.3 DIFC (Distance Image From Fractal Code)
The steps of this method can be explained
as follow.
a. Forming distance and direction images
The distance image A can be formed by rules
bellowed.
A( j , k ) = d i ,
di =
if j = x
(
Ri
(
x
j = 1,2,3, M 1 ,
k = 1,2,3, M 2
−x
)2 + (
y
−y
(c)
(d)
Figure 4: DIFC method. (a) original palmprint image,
(b) distance image, (c) filtered image, and (d) palmprint
feature representation
3.4 AIFC (angle image from fractal code)
The method AIFC has been described in
[2]. The first step of this method is the forming of
angle image A as follows
A( j , k ) = α i , j = 1,2,3, M 1 ,
k = 1,2,3, M 2
)2
(14)
and k = y , otherwise d i = 0 . (15)
Ri
where x D , y D
i
i
)
Di
Ri
Di
α i = arctan
Ri
represent top-left coordinate
of the domain block (see formula (8)) and d i
represent distance between range and domain
277
if
j=x
Ri
yD − yR
xD − xR
and k = y
Ri
, otherwise, α i = 0
(17)
i
(18)
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
(
ISSN: 1992-8645
where x D , y D
i
i
)
www.jatit.org
represent top-left coordinate of
the domain block (see formula (8)) and d i represent
the angle between range and domain block. The
angle image is not represents binary image. The
criterion bellow are added to compute the direction
αi.
if
if
if
if
if
if
x R < x D and y R ≥ y D
x R > x D and y R ≥ y D
x R > x D and y R ≤ y D
x R < x D and y R ≤ y D
x R = x D and y R ≥ y D
x R = x D and y R ≤ y D
then α i = α i
then α i = 180 − α i
then α i = 180 + α i
then α i = 360 − α i
then α i = 90
then α i = 270
The rest of the steps of this method are
similar with the step (b) and (c) in DIFC method.
AIFC method is depend on the selection of
tolerance value (є), filter size (m and n), and M
parameters. The result of each step of FDIFC
method is shown in figure 5.
(a)
E-ISSN: 1817-3195
where size represent the size of range blocks (see
formula (8)), and L represent quadtree
decomposition level. The maximum value (K) of
the level L can be computed as follows.
2 K = min( M 1 , M 2 )
in this case K = 7 because M 1 = M 2 = 128.
Figure 2(b), 4(b), and 5(b) show the
palmprint image representation by using the
criterion with L = K and figure 3(b) is the image
without the criterion.
All methods have been explained above
depend on the selection of tolerance value (є).
Figure 6 show the impact of the tolerance value to
the fractal dimension image representation.
According to the figure, the smaller tolerance value
produced more detail information and the higher
tolerance produced less detail information.
The main advantage of our proposed
methods is the palmprint features can be obtained
directly from the fractal codes palmprint image. Its
mean we can form the palmprint feature directly
from the compressed palmprint image because the
compressed image is collection of fractal codes.
(b)
(a)
(c)
(d)
Figure 5: AIFC method. (a) original palmprint image, (b)
angle image, (c) filtered image, and (d) palmprint feature
representation
3.5 Criterion
The binary image, distance image, and
angle image representation are formed based on the
position of range block. The criterion bellow can be
used to filter the range blocks that will be used to
form those images.
sizei = L − 1
(20)
(19)
(b)
(d)
(e)
Figure 6: Fractal dimension image with various
tolerance є, (a),(b),(c), and (d), represent fractal
dimension image with tolerance 2,3,4, and 5,
respectively.
4.
PALMPRINT FEATURE MATCHING
The degree of similarity between two
palmprint features is computed as follows:
278
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
d rs
www.jatit.org
(xr − xr )(xs − xs )T
(21)
= 1−
[(xr − xr )(xr − xr )T ]12 [(xs − xs )(xs − xs )T ]12
5.
where x r , x s are the mean of palmprint feature x r
and x s respectively. The above equation computes
one minus normalized correlation between
palmprint feature vector x r and x s . The value of d rs
is between 0 – 2. The d rs will be closed to 0 if x r
and x s obtained from two image of the same
palmprint. Otherwise, the d rs will be far from 0.
Figure 7 shows three groups of palmprint
from the same palm and palms with
similar/different line structures and their averages
score are listed in Table 1. The matching score of
group A are close to 0, and the matching score of
group B and C are far from 0. It is easy to
distinguish group A from group B and C using
these scores.
Table 1:Matching Score of groups A, B, and C in figure 7
Metode
FDFC
FDIFC
DIFC
AIFC
Group A
0,1873
0,1093
0,1679
0,1762
Group B
0,6029
0,5462
0,6563
0,5057
Group C
0,7121
0,5763
0,8911
0,6425
(a1)
(a2)
(a3)
Group A: palms from the same person
(b1)
(b2)
(b3)
Group B: palms from different person with similar line
structure
E-ISSN: 1817-3195
EXPERIMENT RESULTS
Those methods have been applied for
palmprint verification system. Verification system
is tested using database of 1050 palmprint images,
are generated from 5 samples from each of the 210
persons randomly selected. The averages of the first
three images from each user were used for training
and the rest were used for testing.
The performance of verification system is obtained
by matching each of testing palmprint images with
all of the training palmprint images in the database.
A matching is noted as a correct matching if the
two palmprint images are from the same palm and
as incorrect if otherwise.
Figure 8 (a) show the probability
distributions of genuine and imposter parts of
FDFC method used tolerance = 3, and feature
vector length = 256 (16 x 16 blocks). The genuine
and imposter parts are estimated from correct and
incorrect matching scores, respectively. Figure 8(b)
show its curve characteristics. The FAR, FRR, and
EER of the system are 0.3226%, 2.0734%, and
1.4523% respectively. We also tested the FDFC
method with tolerance=2.5 and the result show the
FAR=0.2312% and FRR=1.4423%.
Figure 9 (a) show the probability
distributions of genuine and imposter parts of
FDIFC method used tolerance = 3, filter size 5 x 5,
and feature vector length = 256 (16 x 16 blocks).
Figure 9(b) show its curve characteristics. The
FAR, FRR, and EER of the system are 0.344%,
0.479%, and 0.4092% respectively.
FAR and FRR values of FDIFC method in
various filter size and feature vector length with
tolerance = 3 are shown in table 2.
Figure 10 (a) show the probability
distributions of genuine and imposter parts of DIFC
method used tolerance = 3, filter size 5 x 3, and
feature vector length = 256 (16 x 16 blocks). Figure
10(b) show its curve characteristics. The FAR,
FRR, and EER of the system are 0.8445%,
1,7544%, and 1.1164% respectively.
Figure 11 (a) show the probability
distributions of genuine and imposter parts of DIFC
method used tolerance = 3, filter size 5 x 3, and
feature vector length = 256 (16 x 16 blocks). Figure
11(b) show its curve characteristics. The FAR,
FRR, and EER of the system are 0,6743%,
1,7544%, and 1.2759% respectively.
(c1)
(c2)
(c3)
Group 3: palms from different person with different line
structure
Figure 7: Groups of palmprint based on typical of
principal lines
279
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
(a)
(b)
Figure 9: (a) Genuine and impostor distributions of
FDIFC (b) its curve characteristics
(b)
Figure 8: (a) Genuine and impostor distributions of
FDFC, (b) its curve characteristics
(a)
(a)
(b)
Figure 10: (a) Genuine and impostor distributions of
DIFC (b) its curve characteristics
280
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
E-ISSN: 1817-3195
Table 2: FAR and FRR of FDIFC method in various filter
size and feature vector length with tolerance = 3
6.
Feature
Vector
Length
(8 x 8)
FRR
FAR
FRR
FAR
FRR
FAR
0.638
0.609
0.479
0.563
0.479
0.463
(16 x 8)
0.638
0.399
0.479
0.574
0.479
0.489
(16 x 16)
0.638
0.529
0.479
0.504
0.479
0.344
(32 x 16)
0.479
0.675
0.479
0.451
0.479
0.304
(32 x 32)
0.638
0.674
0.479
0.675
0.479
0.344
In this paper, we offer new methods
(FDFC, FDIFC, DIFC and AIFC methods) based
on fractal characteristics to represent the palmprint
features. Those methods have been applied for
palmprint verification system. The experiment
results show that the proposed methods can achieve
an acceptable accuracy rate. The FDIFC method
shows the best performance than other proposed
methods.
The main advantage of using those
methods for feature extraction is the methods can
be formed directly by exploitation fractal codes.
Because of fractal is a great method for image
coding and decoding (image compression), so in
the future, we will applied those proposed methods
for palmprint recognition system from compressed
palmprint images.
Filter 3 x 3
Filter 5 x 3
Filter 5 x 5
CONCLUSIONS AND FUTURE WORK
ACKNOWLEDGEMENT
The authors would like to thank
Directorate General of Higher Education of
Indonesia and Udayana University for supporting
this research.
(a)
REFERENCES:
(b)
Figure 11: (a) Genuine and impostor distributions of
AIFC (b) its curve characteristics
FAR and FRR values of DIFC and AIFC
method in various filter size and feature vector
length with tolerance = 3 are shown in table 3.
Table 3: FAR and FRR of DIFC and AIFC method in
various filter size and feature vector length with
tolerance = 3
DIFC
AIFC
Filter Size/
Number of Blocks
FRR
FAR
FRR
FAR
3 x 3 (8 x 8 block)
2,8708
0,7382
2,5518
0,7877
5 x 3 (8 x 8 block)
2,7113
0,8189
2,8708
0,7683
5 x 5 (8 x 8 block)
2,7113
0,7576
2,8708
0,8199
3 x 3 (16x16 block)
1,7544
0,9222
1,7544
0,8925
5 x 3 (16x16 block)
1,7544
0,8445
1,7544
0,6743
5 x 5 (16x16 block)
1,7544
0,8619
1,7544
0,6998
[1] D. Zhang, W.K. Kong, J.You & M.Wong,
Online Palmprint Identification, IEEE
Transaction on Pattern Analysis and Machine
Intelligence, 25(9), 2003.
[2] Darma Putra, PIFS Code based for Biometric
Palmprint Verification, International Journal
of Computer Science and Information Security
(IJCSIS), Vol. 9, No. 2, 2011, p. 47-52.
[3] T. Yokoyama, K. Sugawara & T. Watanabe,
Similarity-based image retrieval system using
partitioned iterated function system codes, The
8th International Symposium on Artificial Life
and Robotics, Oita, Japan, 2006.
[4] Darma Putra, Erdiawan, High Performance
Palmprint Identification System Based On
Two Dimensional Gabor, Telkomnika UAD,
Vol. 8, No. 3, 2010, p. 195-326
[5] A.K. Jain, Ross A. & S. Pankanti, A Prototype
Hand Geometry-based Verification System,
IEEE Trans. On Circuits and System for Video
Technology, 4(1), 1999.
[6] Darma Putra, Wira Bhuana, Erdiawan,
Pembentukan Kode Telapak Tangan (Palm
Code) berbasis Metode Gabor 2D, Makara UI,
Vol. 15 No. 2, 2011, p. 161-167
281
Journal of Theoretical and Applied Information Technology
th
20 July 2013. Vol. 53 No.2
© 2005 - 2013 JATIT & LLS. All rights reserved.
ISSN: 1992-8645
www.jatit.org
[7] D. Zhang, W. Shu, Two novel characteristics
in palmprint verification: datum point
invariance and line feature matching, pattern
recognition, 32(1), 1999, 691-702.
[8] I Ketut Gede Darma Putra, Sistem Verifikasi
Biometrika Telapak Tangan dengan metode
dimensi fraktal dan lacunarity, Teknologi
Elektro UNUD, Vol. 8 No. 2, 2009, p. 1-6
[9] X.Q. WU, K.Q. Wang & D. Zhang, An
Approach to Line Feature Representation and
Matching for Palmprint Recognition, Journal
of Software, 15(6), 2004.
[10] N. Duta, A.K. Jain & K.V. Mardia, Matching
of Palmprints, Pattern Recognition Letters,
23(1), 2002, 477-485.
[11] Y. Pang, T.B.J. Andrew, N.C.L. David & H.F.
San, Palmprint Verification with Moments,
Journal of WSCG, 12(1-3), 2003.
[12] W. LI, D. Zhang & S. XU, Palmprint
Recognition Based on Fourier Transform,
Journal of Software, 13(5), 2002.
[13] W. Shu & D. Zhang, Automated personal
identification by palmprint, Opt. eng., 37(8),
1998, 2359-2363.
[14] A. Kumar, D.C.M.Wong, C.S. Helen, Anil K.
Jain, Personal Verification using Palmprint
and Hand Geometry Biometric, 2004,
http:/biometrics.cse.msu.
edu/Kumar_AVBPA2003.pdf
[16] T. Connie., A. Teoh, M. Goh & D. Ngo,
Palmprint Recognition with PCA and ICA,
2003,
sprg.massye.ac.nz/ivcnz/proccedings/ivcnz_41
.pdf.
282
E-ISSN: 1817-3195