Palmprint Feature Representation Based On Using Fractal Characteristics.

(1)

(2)

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

(4)

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

(5)

Editorial Board

EDITOR IN CHIEF

Prof. NIAZ AHMAD

FCE, MOE, Islamabad, PAKISTAN

EDITORIAL BOARD

Dr. CHRISTEL BAIER

Faculty of Computer Science, Institute for

Theoretical Computer Science, Technical

University Dresden, GERMANY

Dr. YUSUF PISAN

Department of Software Engineering, Faculty of

Information Technology, University of

Technology, Sydney, AUSTRALIA

.

Dr. YUXIN MAO

School Of Computer & Information Engineering

Zhejiang Gongshang University, CHINA

Dr. MUHAMMAD SHER

Faculty of Basic and Applied Sciences,

Department of Computer Science, International

Islamic University, Islamabad. PAKISTAN.

Dr. ZARINA SHUKUR

Computer Science Dept., Fakulti Teknologi dan

Sains Maklumat, University Kebangsaan

Malaysia, 43600 Bangi, MALAYSIA.

Dr. NOR AZAN MAT ZIN

Department of Information Science, Faculty of

Information Science & Technology, National

University of Malaysia (UKM) 43600 UKM

BANGI, MALAYSIA.

Dr. KHAIRUDDIN BIN OMAR

Faculty of Information Science and

Technology, Universiti Kebangsaan

Malaysia, 43600 Bangi Selangor Darul-Ehsan,

MALYSIA.

Dr. TENGKU MOHD. BIN TENGKU

SEMBOK

Faculty of Information Science and Technology

Universiti Kebangsaan, Malaysia, 43600 Bangi

Selangor Darul-Ehsan, MALYSIA.

Dr PRABHAT K. MAHANTI

Department of Computer Science and Applied

Statistics (CSAS), Hazen Hall Room 311,

University of New Brunswick, Saint John, New

Brunswick, CANADA.

Dr. R. PONALAGUSAMY

Department of Mathematics, National Institute of

Technology, Tiruchirappalli, Tamil Nadu,

INDIA.

Dr. NITIN UPADHYAY

(6)

Group, Birla Institute of Technology and Science

(BITS), Pilani-Goa Campus, NH-17B Bypass

Road, ZuariNagar, Goa, INDIA.

Dr. A. SERMET ANAGN

Eskisehir Osmangazi University, Industrial

Engineering Department, Bademlik Campus,

26030 Eskisehir, TURKEY.

Dr. YACINE LAFIFI

Department of Computer Science, University of

Guelma, BP 401, Guelma 24000, ALGERIA.

Dr. CHRISTOS GRECOS

School of Computing, Engineering and Physical

Sciences, University of Central Lancashire,

Preston PR1 2E, UNITED KINGDOM.

Dr. JAYANTHI RANJAN

Institute of Management Technology, Raj Nagar,

Ghaziabad, Uttar Pradesh, INDIA

Dr. ADEL M. ALIMI

National Engineering School of Sfax (ENIS),

University of SFAX, TUNISIA

Dr. SIKANDAR HAYAT KHIYAL

Department of Computer Science, Fatima Jinnah

Women University, Rawalpindi, PAKISTAN

Dr. ADEL MERABET

Department of Electrical & Computer

Engineering, Dalhousie University, Halifax,

CANADA

DR. HEMRAJ SAINI

CE&IT Department, Higher Institute of

Electronics, Bani Walid. LIBYA

Dr. MAUMITA BHATTACHARYA

SOBIT, Charles Sturt University, Albury - 2640,

NSW, AUSTRALIA

Dr. SEIFEDINE KADRY

Lebanese International University, LEBONON

Dr. AIJUAN DONG

Department of Computer Science, Hood College

Frederick, MD 21701. USA

Dr. S.S.RIAZ AHAMED

Mohamed Sathak Engineering College,

Kilakarai, & Sathak Institute of Technology,

Ramanathapuram , Tamilnadu, INDIA

Dr. ZURIATI AHMAD ZUKARNAIN

University Putra Malaysia, MALAYSIA

Dr. CHELLALI BENACHAIBA

University of Bechar, ALGERIA

Dr. MOHD NAZRI ISMAIL

University of Kuala Lumpur (UniKL)

MALYSIA

Dr. VITUS SAI WA LAM

The University of Hong Kong, CHINA

Dr. WITCHA CHIMPHLEE

Suan Dusit Rajabhat University, Bangkok,

(7)

Dr. SIDDHIVINAYAK KULKARNI

University of Ballarat, Ballarat, AUSTRALIA

Dr. S. KARTHIKEYAN

Caledonian College of Engineering,

OMAN

Dr. DRAGAN R. MILIVOJEVIĆ

Mining and Metallurgy Institute Bor Zeleni

bulevar 35, 19210 Bor, SERBIA

Dr. ABDUL AZIZ

Professor of Computer Science, University of

Central Punjab, PAKISTAN

Dr.P.DANANJAYAN

Professor, Department of ECE, PEC, Puducherry,

INDIA.

Dr. E. SREENIVASA REDDY

Principal - Vasireddy Venkatadri Institute of

Technology, Guntur, A.P., INDIA

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

SHAHBAZ GHAYYUR

Faculty of Basic and Applied Sciences,

Department of Computer Science and Software

Engineering, International Islamic University,

Islamabad. PAKISTAN.

Dr. T.C.MANJUNATH,

Professor & Head of the Dept.,Electronicis &

Communication Engg. Dept,

New Horizon College of

Engg.,Bangalore-560087, Karnataka, INDIA.

Dr. Nacer eddine ZAROUR

LIRE Laboratory, Computer Science

Departement, University Mentouri of

Constantine (UMC)

Dr. RIKTESH SRIVASTAVA

Assistant Professor, Information Systems,

Skyline University College

P O Box 1797, Sharjah, UAE

Dr. Mohd ZAINAL ABIDIN AB

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

(8)

Dr. BONNY BANERJEE

Senior Scientist Audigence, FL, USA, The Ohio

State University, Columbus, OH, USA

Dr. NICKOLAS S. SAPIDIS

Department of Mechanical Engineering,

University of Western Macedonia

Kozani GR-50100, GREECE.

Dr. NAZRI BIN MOHD NAWI

Software Engineering Department, Faculty of

Science Computer Information Technology,

Universiti Tun Hussein Onn

MALAYSIA

Dr. JOHN BABALOLA OLADOSU

Ladoke Akintola University of Technology,

Ogbomoso, NIGERIA

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

Dept. of Computer Science & Engineering, Sri

Sai Madhavi Institute of Science & Technology,

Mallampudi,

Rajahmundry, A.P, INDIA

Dr. KHALID USMANI

Department of Computer Science, Arid

Agriculture University, Rawalpindi, PAKISTAN.

Dr. GUFRAN AHAMD ANSARI

Qassim University, College of Computer

Science, Ministry of Higher Education, Qassim

University, KINGDOM OF SAUDI ARABIA

Dr. Defa Hu

School of Information, Hunan University of

Commerce, Changsha 410205, Hunan, P. R. of

(9)

Home

Journal Rankings

Journal Search

Country Rankings

Country Search

Compare

Map Generator

Help

About Us

Display journal title Just copy the code below and

paste within your html page:

How to cite this website?

F o l l o w u s :

Twitter

SJR is developed by:

J o u r n a l S e a r c h

in Journal Title Search Exact phrase

J o u r n a l o f T h e o r e t i c a l a n d A p p l i e d I n f o r m a t i o n T e c h n o l o g y

Country: Pakistan

Subject Area: Computer Science | Mathematics

Subject Category: Computer Science (miscellaneous) , Theoretical Computer Science

Publisher: Asian Research Publishing Network (ARPN). Publication type: Journals. ISSN: 18173195, 19928645

Coverage: 2010-2013

H Index: 5

Charts Data

Indicators 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 SJR 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,137 0,167

Total Documents 0 0 0 0 0 0 0 0 0 0 0 222 229 622

Total Docs. (3years) 0 0 0 0 0 0 0 0 0 0 0 0 222 451

Total References 0 0 0 0 0 0 0 0 0 0 0 3.880 4.120 9.475

Total Cites (3years) 0 0 0 0 0 0 0 0 0 0 0 0 61 178

Self Cites (3years) 0 0 0 0 0 0 0 0 0 0 0 0 4 17

Citable Docs. (3years) 0 0 0 0 0 0 0 0 0 0 0 0 222 451

Cites / Doc. (4years) 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,27 0,39 Cites / Doc. (3years) 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,27 0,39 Cites / Doc. (2years) 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,27 0,39 References / Doc. 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 17,48 17,99 15,23

Cited Docs. 0 0 0 0 0 0 0 0 0 0 0 0 43 101

Uncited Docs. 0 0 0 0 0 0 0 0 0 0 0 0 179 350

% International

Collaboration 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 5,86 5,68 5,31

(10)

274

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

(11)

Journal of Theoretical and Applied Information Technology

20th July 2013. Vol. 53 No.2

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

275 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)

(g)

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. 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 self-similarity 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 (Ri) and domain block (Di) is noted as

( )

i i i w D

R = (1)

wi is contracted mapping that describes the

similarity relation between Ri and Di, and is usually

defined as an affine transformation as bellow:













+

























=













i i i i i i i i i i i i i i i

o

f

e

z

y

x

s

d

c

b

a

z

y

x

w

0

(2)

where xi and yi represent top-left coordinate of the

Ri , and zi is the brightness value of its block.

Matrix elements ai, bi, ci, and di, are the

parameters of spatial rotations and flips of Di, si is

the contrast scaling and oi is the luminance offset.

Vector elements ei and fi are offset value of space.

We used the size of domain region twice the range size, so the values of ai, bi, ci, and di are 0.5. The

actual fractal code fi bellow is usually used in

practice[3].

(

) (

)

(

D D R R i i i i

)

i

x

y

x

y

size

s

o

f

i i i

i

,

θ

,

=

(3)

where

(

)

i i R

R y

x , and

(

)

i i D

D y

x , represent top-left coordinate position of the range block and domain block, respectively, size is the size of range block,

and

θ

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 counter-clockwise), 6 (rotate 90o clockwise), and 7 (reflection in the line y = -x). The fractal codes of a palmprint image is denoted as follow:



N i i

f

F

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.

( )

R,D ≤ε,

d (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.

(12)

276 The fractal dimension is computed by using the box counting method as bellow [8].

( )

s log

( )

N

( )

s log(s)

D = (6)

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. 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.

( )

j,k c, j 1,2,3, M1,

A = = 

2

, 3 , 2 ,

1 M

k=  (7)

, 1

=

c if

i i andk yR R

x

j= = ,

otherwise c=0 (8)

where A represents the binary palmprint image,

M1 and M2 represent image width and height (in

this paper, M1 = M2 = 128), and

(

)

i i R

R y

x ,

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 = (d1, d2, d3, ..., dM2) (9)

where di represent fractal dimension of ith block.

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.

(a)

(b) (c)

Figure 2: FDFC method. (a) original palmprint image, (b) binary palmprint image, (c) palmprint feature

representation

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 hmxn as follows.

( ) ( ) ( )

x y Ax y h x y

A' , = , ∗ , (10)

where

∗

represent convolution process, and hmxn. 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.

( )

x y

(

A

( )

x y

) ( )

s

AD , =log ' , log (11)

( )

m

n

s

=

max

,

(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

(13)

Journal of Theoretical and Applied Information Technology

20th July 2013. Vol. 53 No.2

277 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:

  



= ,..., ₂

3 , 2 ,

1 _M

v µ µ µ µ (13)

where µ_i represent mean value of block ith.

(a) (b)

(c) (d)

Figure 3: FDIFC method. (a) original palmprint image, (b) binary palmprint image, (c) fractal dimension image,

and (d) palmprint feature representation

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.

( )

j,k d , j 1,2,3, M1,

A = i = 

2 , 3 , 2 , 1 M

k=  (14)

(

) (

2

)

2

i i i

i

i x_R x_D y_R y_D

d = − + −

if i i R y k and R x

j= = , otherwise di =0. (15)

where

(

)

i i D

D y

x , represent top-left coordinate of the domain block (see formula (8)) and di

represent distance between range and domain

block. The distance image is not binary image representation.

b. Filtering the distance images using hmxn

(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

  



= ,..., ₂

3 , 2 ,

1 _M

vJ µ µ µ µ (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.

(a) (b)

(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

( )

j,k ,j 1,2,3, M1,

A =αi = 

2 , 3 , 2 , 1 M

k=  (17)

i R D R D i x x y y − − =arctan α i

iand k yR

R x j

(14)

278 where

(

)

i i D

D y

x , represent top-left coordinate of

the domain block (see formula (8)) and di 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

. i α i i D R D

R x and y y then x

if < ≥ α =α

i i

D R D

if > ≥ α =180−α

i i

D R D

if > ≤ α =180+α

i i

D R D

if < ≤ α =360−α

90 = ≥

= D R D i

if α

270 = ≤

= D R D i

if α

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) (b)

(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.

1 − =L

sizei (19)

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.

) , min( 2 M1 M2

K = ₍₂₀₎

in this case K = 7 because M1 = M2 = 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.

(a) (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

(15)

Journal of Theoretical and Applied Information Technology

20th July 2013. Vol. 53 No.2

279

(

)(

)

(

)(

)

[

]

12

[

(

)(

)

]

12

1 T s s s s T r r r r T s s r r rs x x x x x x x x x x x x d − − − − − − − = (21)

where

x ,

_r

x

_s are the mean of palmprint feature xr

and xs respectively. The above equation computes

one minus normalized correlation between palmprint feature vector xr and xs. The value of drs

is between 0 – 2. The

d

_rswill be closed to 0 if xr

and xs obtained from two image of the same

palmprint. Otherwise, the

d

_rswill 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 Group A Group B Group C

FDFC 0,1873 0,6029 0,7121 FDIFC 0,1093 0,5462 0,5763 DIFC 0,1679 0,6563 0,8911 AIFC 0,1762 0,5057 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

(c1) (c2) (c3)

Group 3: palms from different person with different line structure

Figure 7: Groups of palmprint based on typical of principal lines

5. 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.

(16)

280

(a)

(b)

Figure 8: (a) Genuine and impostor distributions of FDFC, (b) its curve characteristics

(a)

(b)

Figure 9: (a) Genuine and impostor distributions of FDIFC (b) its curve characteristics

(a)

(b)

(17)

Journal of Theoretical and Applied Information Technology

20th July 2013. Vol. 53 No.2

281

Table 2: FAR and FRR of FDIFC method in various filter size and feature vector length with tolerance = 3

Feature Vector Length

Filter 3 x 3 Filter 5 x 3 Filter 5 x 5

FRR FAR FRR FAR FRR FAR

(8 x 8) 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

(a)

(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

Filter Size/ Number of Blocks

DIFC AIFC

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

6. CONCLUSIONS AND FUTURE WORK

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.

ACKNOWLEDGEMENT

The authors would like to thank Directorate General of Higher Education of Indonesia and Udayana University for supporting this research.

REFERENCES:

[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,

(18)

282 [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,

Palmprint Feature Representation Based On Using Fractal Characteristics.

Table of Contents

Energy-Aware Node Placement in Wireless Sensor Network Using ACO

Indexing and Abstracting Information

Editorial Board

Prof. NIAZ AHMAD

EDITORIAL BOARD

Dr. CHRISTEL BAIER

Dr. MUHAMMAD SHER

BANGI, MALAYSIA.

SEMBOK

Dr PRABHAT K. MAHANTI

Dr. A. SERMET ANAGN

Dr. JAYANTHI RANJAN

CANADA

DR. HEMRAJ SAINI

Dr. MAUMITA BHATTACHARYA

Dr. ZURIATI AHMAD ZUKARNAIN

Dr. SIDDHIVINAYAK KULKARNI

INDIA.

KHAMITKAR

Dr. M. IQBAL SARIPAN

SHAHBAZ GHAYYUR

Dr. T.C.MANJUNATH,

Dr. Mohd ZAINAL ABIDIN AB

Dr. BONNY BANERJEE

MALAYSIA

Dr. ASHRAF IMAM

Dr. KHALID USMANI

PALMPRINT FEATURE REPRESENTATION

Keywords: Biometrics, Fractal Codes, Fractal Dimension, Feature Extraction, Palmprint Recognition.

1. INTRODUCTION

Journal of Theoretical and Applied Information Technology

2. FRACTAL CHARACTERISTICS

2.1 Self-Similarity

0

0

0

0

3. PALMPRINT FEATURE REPRESENTATION

3.1 FDFC (fractal dimension from fractal codes)

3.2 FDIFC (Fractal Dimension Image From Fractal Codes)

Journal of Theoretical and Applied Information Technology

3.3 DIFC (Distance Image From Fractal Code)

3.4 AIFC (angle image from fractal code)

3.5 Criterion

4. PALMPRINT FEATURE MATCHING

Journal of Theoretical and Applied Information Technology

Metode Group A Group B Group C

5. EXPERIMENT RESULTS

Journal of Theoretical and Applied Information Technology

DIFC AIFC

6. CONCLUSIONS AND FUTURE WORK

REFERENCES: