Review On:-Finger Vein Detection using Repeated Line

ISSN-2349-1841(Online)
Volume 1, Issue 1, January 2015
International Journal of Research Development
& Innovation (IJRDI)
Research paper
Available online at: www.ijrdi.com
Review On:-Finger Vein Detection using Repeated
Line, Even Gabor and Multilinear Discriminant
Analysis(MDA)
Manpreet Kaur#1, Geetanjali Babbar#2
#
Mtech student, CEC LANDRAN, PTU
1
[email protected]
2
Assistant Professor
CEC LANDRAN, Punjab Technical University
2
[email protected]
Abstract— As a new biometric technique, finger vein detection
has attracted lots of attentions and used in wide range of
applications. Finger vein detection is a physiological
characteristics-based biometric technique, and it uses vein
patterns of human finger to perform identity authentication.
Vein patters are network of blood vessels under a person’s
skin. Even in the case of identical twins the finger-vein
patterns are believed to be quite unique [1].This make finger
vein detection secure biometric for individual identification.
Many techniques and methods are available for finger vein
detection. Although these methods produce efficient results
but have certain limitations. In this paper, following different
techniques are discussed and analyzed i.e., Repeated Line
Tracking, Gabor Filter and Multilinear discriminant
analysis(MDA) to reduce the limitations of previous methods
.The work focuses on study and performance measuring
analyzes of these techniques using MATLAB.
Keywords:- Finger vein, biometric system, Repeated Line
tracking,
Gabor
filter,
Multilinear
discriminant
analysis(MDA).
I. INTRODUCTION
Personal identification technology is applied to a wide range of
systems including controlling access, computer login, and ecommerce. Biometrics is the statistical measurement of human
physiological traits. Biometric techniques for person detection
have been attracting attention these days because conventional
means like keys, passwords, and security PIN numbers have
All Rights Reserved
problems like theft, loss, and reliance on the user’s memory. In
the area of biometric identification, security and convenience of
the system are important [1]. In particular, the systems require
high accuracy and fast response times. Biometric methods
include those based on the pattern of fingerprints [9– 11], facial
features, the iris, the voice, the hand geometry, or the veins on
the back of the hand. However, these methods do not
necessarily ensure confidentiality because the features used in
the methods are exposed outside the human body. This result in
forgery in these methods. To reduce these issues, biometric
system using patterns of veins within a finger, that is, patterns
inside the human body is proposed [2,3].
In embryology, the formation of vascular tissues is subject to the
combined action of both deterministic and
random processes in embryogenesis [13,14]. This demonstrates
that the uniqueness of finger-vein pattern essentially comes
from the randomness of vein networks.
The finger veins used for biometrics are a kind of superficial
veins that are usually present between the two layers of
superficial fascia and not accompanied by arteries [11]. As the
veins network course in the subcutaneous tissue, the visible light
usually not capable of imaging them correctly though some of
them can be directly seen under the skin. So, to visualize veins
in the human tissues, the near infrared (NIR) light (700–900 nm)
is often used in practical applications because they can penetrate
the human skin and be absorbed greatly by the
deoxyhemoglobin [12].
P a g e | 50
Manpreet Kaur et.al...
www.ijrdi.com
International Journal of Research Development & Innovation (IJRDI)
Volume 1, Issue 1, January 2015, Pg. 50-54
if they are from the same finger of same hand and of the same
person. Two literatures mentioned are the foundation of finger
vein recognition, which open the era of finger vein recognition.
(a) Imaging device (1)
(b) Imaging device(2)
(c)Finger vein image (1) (d) Finger vein image (2)
Fig.1 Prototype of finger-vein imaging device (a, b) and examples of infrared
images of a finger (c, d)
Finger Vein proved that each finger has unique vein patterns so
that it can be used in personal verification. To get the vein
pattern, a person
inserts a finger into an attester terminal
containing a near-infrared LED light and a monochrome CCD
camera as shown in fig.1. The hemoglobin in the blood, absorbs
the closer infrared light rays, the pattern of veins in the palm
side of the hand of finger is captured as a pattern of shadows,
which makes the vein system. The camera records the digitized
raw data and the image, and sent it to a database of registered
images. The finger is scanned as before and the data is sent to
the database of registered images for authentication purposes
which takes less than two seconds.
A typical finger vein identification system mainly includes
image acquisition, preprocessing, feature extraction and vein
matching, as shown in Fig.2. In the finger vein image
acquisition process, the light intensity has a great impact on the
quality of finger vein image. The Stronger light can cause the
overall image bright and make the vein disappear. The lower
light can cause the overall image dim and result in very little
difference between the blood vessels and the background. In
order to get sufficient extraction of the finger vein feature, the
image requires preprocessing. The preprocessing is mainly
divided into image normalization and image Enhancement.
Methods of feature extraction mainly include: threshold
segmentation methods, the maximum principal curvature
methods.
Kono et al [3], Japanese medical researchers, proposed finger
vein based identity detection and gave more effective feature
extraction method. Yanagawaetal [4] proved the diversity of
human finger vein patterns and the usefulness of finger veins for
identity identification on 2, 024 fingers of 506 persons. They
state that, two finger vein patterns are identical only in that case
All Rights Reserved
A. Why we use finger vein?
As a biometric characteristic, finger vein has various desirable
properties, such as universality, permanence, distinctiveness and
acceptability. In addition to this as compared with other
biometric characteristics (for example, face, fingerprint and so
on) it has following two advantages :
1. Living body identification: means that only vein in finger
are captured and further used to perform identification.
2. Internal characteristic. It is hard to copy or forge finger vein,
and very little external factor can damage finger vein, which
guarantee the high security of finger vein recognition.
These above mentioned two advantages make finger vein an
irreplaceable biometric traits, and gain attentions from more
and more research teams. This make it fairly recent
technological advance in the field of biometrics that is being
applied to different fields such as financial, medical, law
enforcement facilities and other applications where high levels
of security or privacy is very important.
B. Application of finger veins
Various applications of finger vein detection are:
1. In the field of forensic science and crime scene
investigations: Vascular technology means vein
matching, is a technique of biometric identification
through the analysis of the patterns of vessels visible
from the surface of the skin. Though used by the
Central Intelligence Agency and Federal Bureau of
Investigation , the method of identification is still in
development. This can also be used in conjunction with
existing forensic data in support of a conclusion [4].
Criminals identification based on finger vein will be
more reliable as compared to finger prints.
2. Employee time and attendance tracking: This
innovative reader eliminates buddy punching and
increases productivity without the need for ID cards.
This make hardware configuration scalable from one
reader to multiple readers for attendance and access
control.
3. Controlling Access: The finger vein authentication
system prevents leakage of corporate information and
entry of unidentified persons into home or office
buildings. This system can also be utilized with
combination of corporate identification cards or antitheft monitors to achieve multiple layers of security
control.
P a g e | 51
Manpreet Kaur et.al...
www.ijrdi.com
International Journal of Research Development & Innovation (IJRDI)
Volume 1, Issue 1, January 2015, Pg. 50-54
4.
Protecting Computer Login: PC users can now guard
against illegitimate access or leakage of information on
their computers by using the vein patterns of their
fingers as their PC login. As this method is free from
any forgery, results in better protection. This finger
vein authentication technology can also centralize
information scattered in different locations to enable
more effective information management.
5. Revolutionizing Banking: Vein pattern recognition
technology eliminates the need for bank cards or
security PIN numbers , thus overcome problems
relating to loss, theft or duplicity of cards and
passwords. Banks may also use finger vein
authentication systems to effectively manage
transactions at bank counters and depositories; hence
reduce case of thefts like frauds.
6. Designing more secure safe : Safe accepting finger
veins is more reliable as compare to safes with finger
prints hence more efficient in protection of confidential
files and data. Since finger vein does not suffer from
any forgery related problems. Such safes also used for
protecting military weapons.
II. BLOCK DIAGRAM OF PROPOSED WORK
The block diagram of the proposed system is shown in Figure
Load
Image
Preproce
ssing
Feature
extraction
Image Normalization
Matchi
ng
1. Binarization / Normalization: Each of the acquired fingervein images is first subjected to normalization, in which
binarization is done[1][5].A binary image is a digital image that
has only two possible values for each pixel.
we convert grey scale image to black and white i.e in a binarized
form 1 and 0.
2. Edge detection: In this step edges of a finger vein image are
detected using soble edge detector so that we extract finger vein
area properly.
3. Vein ROI: ROI(REGION OF INTEREST) is used to acquire
important region of a finger vein image, due to which less time
is consumed and unwanted area is excluded.
4. Image Enhancement: .it is used to improve the
quality .Image got damaged or its quality decrease during the
process of clicking the image or transferring the image then the
image enhancement is use to improve quality of finger vein
image .it improves the color ,contrast, brightness and moreover
reduces noise present in an image[14].
Feature Extraction: - Transforming the input data into the set
of features is called feature extraction. The feature extraction is
done to extract the thickness, length and shape of the vein
pattern. The three methods are used to extract the features are
Repeated Line Tracking, Even Gabor filter and MDA.
Database
Edge Detection
Result
Analysis
ROI Extraction
Image Enhancement
Fig 2: Block diagram of proposed system
Image Acquisition:- Image acquisition means acquiring an
image from the source stored in some hardware or from direct
source. In image acquisition the browse image is loaded for
further processing.
All Rights Reserved
Preprocessing:- After acquiring finger vein image
preprocessing of an input image is done in second step which
includes;
Finger Vein Matching: - The output come from all the previous
steps will be matched with the data stored in the database, which
shows whether the acquired vein is matched or not. Here SURF
(speed up robust feature) is used for matching vein pattern. The
SURF has better speed and accuracy as it uses object
reorganization feature. The database consists of the features of
all vein images [14, 16].
III. LITERATURE REVIEW
Ajay Kumar et al. proposed a method for Human
Identification Using Finger Images. They proposed a new
approach to improve the performance of finger-vein
identification systems presented in the literature. The proposed
system simultaneously capture the finger-vein and low
resolution fingerprint images and then combines these two
evidences using a novel score-level combination strategy. The
utility of low-resolution fingerprint images acquired from a
webcam is examined to ascertain the matching performance
from such images. They develop and investigate two new score-
P a g e | 52
Manpreet Kaur et.al...
www.ijrdi.com
International Journal of Research Development & Innovation (IJRDI)
Volume 1, Issue 1, January 2015, Pg. 50-54
level combinations, i.e., nonlinear fusion and holistic, and
comparatively evaluate them with more popular score-level
fusion approaches to ascertain their effectiveness in the
proposed system[1].
Pengfei Yu et al. proposed a method Fingerprint Image
Preprocessing Based on Whole-Hand Image Captured by
Digital Camera. They present, some fingerprint image preprocessing approaches based on a whole-hand image captured
by digital camera. The pre-processing methods include 1.key
point location 2.finger image segmentation 3.fingerprint region
extraction. At first, the key points including fingertips and
valley points, that are key points, are located from the hand
contour image. In Second step, the middle finger is cropped
from the hand image based on the information of the key points'
positions and the knuckle's texture near the finger root. At last,
the fingerprint is extracted from the middle finger through the
first knuckle's texture. Due to , low resolution of fingerprint
images, linear projection methods like Principle Component
Analysis (PCA) and Linear Discriminant Analysis (LDA) are
used for fingerprint feature extraction. Experimental outcomes
on a database of 86 hands (10 impressions per hand) show that
these approaches are effective and more accurate [2].
Miura, N., Nagasaka et al. : propose a method of personal
identification based on finger-vein patterns. In this paper
they proposed a technique in which an image of a finger
captured under infrared light contains not only the vein pattern
along with irregular shading produced by the various
thicknesses of the finger bones and muscles. This method
extracts the finger-vein pattern from the unclear image by using
line tracking method that starts from various positions.
Experimental results show that, it achieves robust pattern
extraction. The equal error rate was 0.145% in personal
identification [3].
Jinfeng Yang et al. proposed a technique for an improved
method for finger-vein image enhancement. They found an
improved
approach,
incorporating
both,
directional
decomposition and Frangi filtering, to enhance finger veins.
Firstly, background subtraction is processed, for improving
image contrast and eliminates uneven illumination. Secondly,
the vein information as ridge-like directional characters in
image is detected by directional filter bank and enhanced
through Frangi filtering after noise removal. At end the
enhanced images are obtained from enhanced directional images
with help of a fine reconstruction rule. Experimental results
state that, the new proposed approach is more robust, effective
and accurate in enhancing finger-vein images [14].
All Rights Reserved
IV.
DIFFERENT TECHNIQUES OF FEATURE
EXTRACTION
Various techniques used in proposed work are:
1) Repeated line Tracking:
Repeated line tracking method gives a promising result in
finger-vein identification [1][3]. Line tracking starts at different
positions, and moves along the direction of vein pattern pixel by
pixel. The idea is to trace the veins in the image by chosen
directions according to predefined probability in the horizontal
and vertical orientations, and the starting position, that is, seed is
randomly selected; the whole process is repeatedly done for a
certain number of times [3].
2) Even Gabor filter:
Gabor filter, is a linear filter used for edge detection. Its
frequency orientation representation is similar to those of the
human visual system; and they have been found to be
particularly appropriate for texture representation and
discrimination. Hence in spatial domain; a 2D Gabor filter is a
Gaussian kernel function modulated by a sinusoidal plane wave
[1][5]. Gabor filters: are self-similar, which means, all filters
can be generated from one mother wavelet by dilation and
rotation [10].Its impulse response is defined by a Gaussian
function multiplied by a harmonic function. Because of
(Convolution theorem) the multiplication-convolution property,
the Fourier transform of a Gabor filter's impulse response is the
convolution of the Fourier transform of the Gaussian function
and Fourier transform of the harmonic function [15]. These
filters have been shown to posses optimal localization properties
in both frequency and spatial domain. And thus are well suited
for texture segmentation problems. Gabor filter can be viewed
as a sinusoidal plane of particular frequency and orientation [1],
modulated by a Gaussian envelope.
G(x, y) = s(x, y) g(x, y)
Where g(x, y) is 2D Gaussian envelope and (x, y) is complex
sinusoid
Formula:
′2
𝑔 = 𝑒𝑥𝑝 (\𝑓𝑟𝑎𝑐{𝑥 ′2 + 𝛾 2𝑦 }{2𝜎 2 }) 𝑛𝑔
\𝑒𝑥𝑝(𝑖(2𝜋\𝑓𝑟𝑎𝑐{𝑥 ′ }{𝜆} + 𝜓))
3) Multi-linear Discriminant Analysis (MDA):
It is Dimensionality Reduction-based Methods in which
multiple interrelated subspaces can collaborate to discriminate
different classes. Subspace learning methods usually transform
image into low-dimensional space to classify.
In
P a g e | 53
Manpreet Kaur et.al...
www.ijrdi.com
International Journal of Research Development & Innovation (IJRDI)
Volume 1, Issue 1, January 2015, Pg. 50-54
transformation, they keep discriminating information and
remove noises[5].
[6]Prabjot Kaur et al., “Comparative Study on Human Identification for Finger
Vein Images” International Journal of Latest Scientific Research and
Technology 1(2), pp. 37-40 ISSN: 2348-9464, July – 2014.
Various advantages of MDA:
1) The MDA algorithm can avoid the curse of dimensionality
and serves the small sample size problem.
2) The computational cost in the learning stage is reduced.
3) The extension from vector to tensor for data representation
and feature extraction is general and it can also be applied on
SVM and many other algorithms to improve algorithm learn
ability and effectiveness.
[7]Naoto Miura, Akio Nagasaka, Takafumi Miyatake “Feature Extraction of
Finger-Vein Patterns Based on Repeated Line Tracking and its Application to
Personal Identification” Published by Springer on Machine Vision and
Applications”, Volume 15, Issue 4, pp 194-203, October 2004.
V.
CONCLUSION
To implement new algorithm for the finger-vein detection,
which is more reliable in extracting the finger-vein features and
result in higher accuracy than the previously proposed fingervein detection approaches. Numbers of Finger Vein detection
techniques have been proposed earlier but they were not secure
enough and have some limitations.
[10]Naoto Miura, Akio Nagasaka, Takafumi Miyatake “Extraction of FingerVein Patterns Using Maximum Curvature Points in Image Profiles Published by
IEEE in 2005.
Finger Vein image detection using Repeated Line Tracking or
Gabor Filter alone could not provide better results. New fingervein matching scheme is based on finger vein processing using
three methods i.e. Repeated line tracking, Even Gabor filter and
Multi-linear Discriminant Analysis(MDA) which results in
better Average Recognition Performance and PSNR value ,as
compared to previous algorithms and works more effectively in
more realistic situations and turned to a more accurate
performance.
REFERENCES
[1] Ajay Kumar,Senior Member, IEEE, and Yingbo Zhou; “Human
Identification Using Finger Images” IEEE transactions on image processing,
vol. 21, no. 4, april 2012.
[2] Pengfei Yu, Dan Xu , Haiyan Li , Hao Zhou; “Fingerprint Image
Preprocessing Based on Whole-Hand Image Captured by Digital Camera”.
International conference
on Computational Intelligence and Software
Engineering2009. CiSE 2009.
[8].David Mulyono & Horng Shi Jinn “A Study of Finger Vein Biometric for
Personal Identification” Published by IEEE in 2008.
[9]Kejun Wang, Hui Ma, Oluwatoyin P. Popoola and Jingyu Li “Finger Vein
Recognition” Published by ISSN in 2008.
[11]Zharov V, Ferguson S, Eidt J, Howard P, Fink L, Waner M(2004) “Infrared
imaging of subcutaneous veins”. Lasers Surg Med 34(1):56–61
[12]Guidolin D, Nico B, Mazzocchi G, Vacca A, Nussdorfer GG, Ribatti D
“Order and disorder in the vascular network.Leukemia” Published by IEEE in
2004
[13] Jinfeng Yan, Minfu Yan; “An improved method for finger-vein image
enhancement. Signal Processing”(ICSP), 2010 IEEE 10th International
Conference.
[14] Yang JF, Yang JL, Shi YH “Finger-vein segmentation based on multichannel even-symmetric Gabor filters” Published by IEEE international
conference on intelligent computing and intelligent systems in 2009.
[15] David Mulyono & Horng Shi Jinn “A Study of Finger Vein Biometric for
Personal Identification” Published by IEEE in 2008 .
[16] Kejun Wang, Hui Ma, Oluwatoyin P. Popoola and Jingyu Li “Finger Vein
Recognition” Published by ISSN in 2008.
[17] Naoto Miura , Akio Nagasaka , Takafumi Miyatake “Extraction of FingerVein Patterns Using Maximum Curvature Points in Image Profiles” Published
by IEEE in 2005.
[3] Miura, N., Nagasaka, Akio Nagasaka, Takafumi Miyatake:“Feature
extraction of finger-vein pattern based on repeated line tracking and its
application to personal identification”. Machine Vision andApplications”, 21
July 2004 –_c Springer-Verlag 2004.
[4]Yanagawa, T., Aoki, S., Ohyama, T.: “Human finger vein images are diverse
and its patterns are useful for personal identification,” MHF Preprint Series,
Kyushu University, pages 1–7, 2007.
[5]Shuicheng Yan, Member, IEEE, Dong Xu, Multilinear Discriminant Analysis
for Face Recognition IEEE transactions on image processing, vol. 16, no. 1,
january 2007
All Rights Reserved
P a g e | 54