Jammi Ashok et. al.

/ (IJCSE) International Journal on Computer Science and Engineering

Vol. 02, No. 07, 2010, 2402-2408

AN OVERVIEW OF BIOMETRICS
JAMMI ASHOK 1 VAKA SHIVASHANKAR 2 P.V.G.S.MUDIRAJ 3
Professor and Head Assistant Professor Associate Professor,MCA Dept.
Department of IT, Department of MCA Adams Engg. College,
GCET, Hyderabad, India. KCCS, Warangal Paloncha,Khammam.

Abstract 2. TYPES OF BIOMETRICS

The term biometrics is derived from the Greek Based on the above guidelines, several biometrics are
words bio meaning “life” and metrics meaning “ to measure” . being developed and are in use .This paper describes
Biometrics refers to the identification or verification of a person popularly used biometrics in terms of the character measured
based on his/her physiological and/or behavioral characteristics. , the devices used to collect the biometric , features extracted
Several verification/identification based biometrics have , the algorithms used and the areas of applicability.
evolved based on various unique aspects of human body, ease
of acquiring the biometric, public acceptance and the degree of 2.1 FINGERPRINT BIOMETRIC METHOD
security required. This paper presents an overview of various
biometrics in use/proposed and their applicability to different Fingerprint identification is popular because of their ease
activities. in acquisition, number of sources (ten fingers) and their
acceptance over a long time by law enforcement offices.
Keywords: Biometrics, FMR, FNMR, ROC, CMC Fingerprints form part of an individual’s phenotype and are
not determined by genetics and hence qualify as good
1. INTRODUCTION biometric. A fingerprint appears as a series of ridges with
pores (sweat glands ) and valleys between these ridges as
1.1 NEED FOR BIOMETRICS shown in fig1. In a fingerprint a minutia is a point where a
In the present times, when most transactions - ridge ends or splits. A typical finger would have 30 to 60
financial or otherwise - are automated and many of them minutia points. Minutia is the feature which is extracted for
networked, security has emerged as a most important issue . fingerprint biometric .
Security is usually in the form of possessions ( like ID cards ,
keys ) or secret knowledge ( like password , PIN ) . This type of
security is not failsafe as for example, ID cards may be lost ;
passwords may be forgotten or compromised [1 ] . A strong
need was thus felt for more robust authentication methods and
extensive research ensued in this area. This led to the concept
of using human body parts or human mannerism itself as
security and authentication measure, and finally to the
emergence of biometrics as a field by itself . It is now widely
accepted that any positive identification of a person must
include biometric identification [1 ] .

1.2. PREREQUISITES OF A GOOD BIOMETRIC

Any aspect of human physiology or behavior that Fig1: A typical minutia map
can be accepted as a biometric should satisfy five properties
described by Clarke which are as follows : Four technologies are in use to extract fingerprint images.
a) Universality : Every person should have the biometric These are as listed below
characteristic.
b) Uniqueness : No two persons should be the same in terms of a)Optical Sensors : These sensors capture visual image of
the biometric characteristic finger surface. Finger touches the surface of a prism and
c)Permanence: The biometric characteristic should be invariant LEDs provide a light source. Image is captured after its total
over time. internal reflection in the prism, by a Charge Coupled Device
d)Collectability: The biometric characteristic should be IC (CCD-IC) or CMOS Camera. Optical sensors are reliable
measurable with some practical sensing device. and inexpensive. However they are bulky and prone to
e)Acceptability: The public should have no strong objection to surface dirt and dust that affects the quality of fingerprint
the measuring or collection of the biometric. collected.

ISSN : 0975-3397 2402

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

b) Capacitive Sensors : These sensors scan surface of finger used for matching. The advantage with this method is the
using dielectric measurements to distinguish ridges and small template size typical space required being < 400 bytes
valleys. Higher dielectric constant of ridges results in higher per finger . Due to the small template size, matching becomes
capacitance than that of valleys which contain air. faster .However the minutiae extraction process itself takes
Capacitive sensors produce better image quality over wider a long time.
operating conditions. However they are expensive ,consume In image based matching, the location , orientation as well
more power and also do not work well with dry fingers . as a portion of the image around the minutia point is stored .
c) Thermal sensors : These sensors consist of contiguous Patches of reference image are placed on test image; each
arrangement of heating elements and thermal sensors and patch is shifted and rotated over the test image to find a best
capture images based on differentials in heat emission fit .Once all patches are aligned to their best spots , the
between the ridges and valleys. Heat map is converted to an locations are used to verify the relative distance between
optical image of ridges which are cooler due to presence of patches . Enough patches at the right places indicate a match.
sweat pores and valleys which are warmer. Thermal sensors Notable fingerprint biometric systems are i) IAFIS (
are compact and inexpensive. But they consume more power Integrated Automatic Fingerprint Identification System)
and are ineffective on warm days. maintained by FBI and ii) US-VISIT which confirms
d) Radio Frequency Sensor : These sensors scan sub- whether a person applying for entry/exit from USA is the
surface to get a true image of the finger. They use reflected same person as the one granted visa by the department of
RF beam to create an image of the layer. RF sensors are not state.
affected by dirt or other impurities , have improved accuracy
and reliability .They are also robust and small in size. 2.2 FACE RECOGNITION
Also, it is very difficult to fake the finger with this sensor as Face appearance is a biometric which is used everyday
it takes subsurface image . by everyone as a primary means of recognizing other
humans[1]. Because of this naturalness it is more acceptable
Sensor Measu Advant Drawbacks Special than other biometrics. Face image acquisition is done in the
used res ages Feature following ways
a) Single image : This consists of digital photographs
obtained using cameras or scanners.
Optical Visual Reliabl Affected
sensor image e, by dust, b) Video Sequence : This is obtained from surveillance
inexpen dirt cameras. However, due to low spatial resolution, it is not
sive very useful for face recognition.
Capaciti Dielec Better Expensive,
ve tric image consume c) 3D Images : This is based on skin/skull geometry and
sensor consta quality more requires 3D images of the face instead of 2D images .
nt power, not Newer face recognition techniques such as Stereo ,
good on structured light and phase based ranging are used for
dry fingers capturing 3D images .[1]
Therma Heat Compa Consume Not
l Sensor emissi ct, more effective Face recognition approaches in turn are divided into two
on inexpen Power on categories
sive warm
days a) Face appearance based : Here, a face image is
transformed into what is known as Eigen faces .To
Radio Reflec High Scan generate a set of eigenfaces , a large set of digitized
Frequen ted RF accurac subsurfac images of human faces taken under the same lighting
cy beam y, e, conditions are normalized to line up the eyes and
Sensor compac difficult mouths. They are then resampled at the same pixel
t to fake resolution . Eigenfaces can be extracted out of the image
by means of a mathematical tool called PCA ( Principal
Component Analysis ).

b) Face geometry based : This is based on face features.

Table 1 : comparison of various fingerprint sensors Features like the rim of the nose and cheeks of the
subject are detected and their geometric relationships are
Fingerprint identification is done using minutiae used for recognition of the face as in fig 2.
based matching or pattern matching . In minutiae based
matching , the location and orientation of minutiae points are

ISSN : 0975-3397 2403

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

derived from the immediately surrounding eye region .

[16]
Iris recognition based on John Daugman’s algorithms
, is used by the United Arab Emirates (UAE) Ministry of
Interior for recognizing foreigners entering the UAE , at 35
air, land, and sea ports. Each traveler is compared against
about a million Iris codes on a watch-list through internet
Fig 2 :Geometry based face recognition [16] links ; the time required for an exhaustive search through the
database is about 1 second. On an average day, about 12,000
arriving passengers are compared against the entire watch-
list i.e. about 12 billion comparisons per day . So far, about
7,500,000 exhaustive searches against that database have
been done, making about 7 trillion iris comparisons
altogether. A total of 73,180 matches have so far been found
between persons seeking re-entry into the UAE and persona
non grata on the watch list.[4]

2.4 HAND RECOGNITION

In hand recognition, the geometric features of the hand
Fig 3 : Iris image [17] such as the lengths of fingers and the width of the hand are
measured using a charged couple device camera (CCD) and
various reflectors and mirrors. Black and white pictures of i)
An image of the top of the hand ; and ii) An image of the
side of the hand are captured. Unique features in the structure
of the hand such as finger thickness, length and width, the
distances between finger joints, the hand’s overall bone
structure, etc are also recorded. To enroll, the user places his
or her hand onto a platen three different times ; three images
Fig 4 :Hand geometry acquisition device sensing the top and side of are captured and averaged . The resulting image forms the
hand[18] basis for the enrolment template, which is then stored in the
database of the hand geometry scanner. The enrolment phase
Face biometric is implemented by Queensland Transport , can be completed within five seconds . In the verification
Australia for Driver’s License with technical support from phase, the user is prompted to place his/her hand only once
Unisys Corporation, Sydney. [9] on the platen. An image is captured, and forms the basis for
the verification template. The verification template is
2.3 IRIS RECOGNITION compared against the enrolment template, in the same
The colored part of the eye between he pupil and sclera fashion as fingerprint recognition. The verification phase can
is called the Iris. Since Iris is a protected internal organ with be accomplished in just under one second .This technology is
a complex texture , and because it is unique from person to mostly used in physical access entry applications. [5]
person and stable throughout life, it forms a very good
biometric. 2.5 ADDITIONAL BIOMETRICS
Iris image acquisition is done in two ways
a) Retina scan
a) Daugman System : In this system, an LED based point The retina biometric analyzes the layer of blood vessels
light source is used along with a standard video camera. located at the back of the eye. This technique uses a low-
The system captures images with the Iris diameter intensity light source through an optical coupler and scans
between 100 to 200 pixels from a distance of 15 to 46cm the unique patterns of the retina’s blood vessels . Retina
using 330mm lens. John Daugman at the university of scanning is quite accurate and very unique to each individual
Cambridge computer laboratory developed Gabor similar to the Iris scan; but unlike the Iris scan, it requires
wavelet based Iris recognition algorithm which is the the user to look into a receptacle and focus on a given point
basis for almost all commercially available Iris for the user's retina to be scanned. This is inconvenient for
recognition systems people who wear glasses and those concerned about close
contact with the scanning device. This technique is more
b) Wilde’s system : This system images the Iris with intrusive than other biometric techniques although the
approximately 256 pixels across the diameter from 20cm technology itself is very accurate for use in identification,
using 80 mm lens and is area based i.e. it captures the verification and authentication.[9] Additionally, diseases
iris as part of a larger image which also contains data

ISSN : 0975-3397 2404

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

such as cataracts can cause the retina to change making this to be promising, no commercial systems are
technique not reliable over a period of time.[11] available.[8]
b) Vein scan biometric : Vein scan biometric technology
identifies a person from the patterns of the blood vessels 3. PERFORMANCE METRICS
in the back of the hand. The technology uses near-
infrared light to detect vein vessel patterns. Vein While considering performance measures Biometric
patterns are distinctive between twins and even between systems are classified into verification systems in which a
a person's left and right hand. These are developed biometric matcher makes a 1:1 match decision based on a
before birth, highly stable and change through one's life score s ; and identification systems which makes a 1:m
only in overall size. The technology is not intrusive, and match decision .
works even if the hand is not clean. It is commercially
available and implemented by Fijitsu of Japan[8] 3.1 PERFORMANCE METRICS FOR VERIFICATION
SYSTEMS :
c) Facial thermograph
3.1.1 False Match Rate (FMR) & False Non Match Rate
Facial thermograph detects heat patterns created by the (FNMR)
branching of blood vessels and emitted from the skin. β and β’ be two real world biometrics - say two
An infrared camera is used to capture the resulting fingerprints or two faces ; B and B’ being the associated
images. The advantages of facial thermograph over other machine representations of these biometrics. Then B = f(β)
biometric technologies are - it is not intrusive, no and B’ = f(β’) where ‘f’ represents the feature extraction
physical contact is required, every living person process. However in the real world β and β’ are functions of
presents a usable image, and the image can be collected time. Also ‘f’ is a function of duration of sensing and other
on the fly. Also, unlike visible light systems, infrared environmental factors .This variability is reflected by writing
systems work accurately even in dim light or total the above equations as B = B(t) = ft(β(t))
darkness. Although identification systems using facial B’ = B’ (t’) = ft’(β’(t’))
thermograms were undertaken in 1997, the effort was Biometric match engines , make a decision by computing a
suspended because of the cost of manufacturing the measure of the likelihood that the two input samples subject1
system. and subject2 are the same (but at different times) and
d) Skin pattern therefore that, subject1 (at time t) and subject2 (at time t’)
The exact composition of all the skin elements such as are same in real world. This measure called the similarity
skin layer thickness, undulations between layers , measure takes β and β’ as inputs and computes a score
pigmentation, collagen fibers etc is distinctive to each s(B’,B)=s(B’(t’),B(t))= s[ft’(β’(t’)),ft (β(t)) ]
person. Skin pattern recognition technology measures Where B is the enrolled sample’s template at time t and is
the characteristic spectrum of an individual's skin. A rarely changed and B’ is the live query sample’s template.
light sensor illuminates a small patch of skin with a Higher is the value of s greater is the likeliness that the two
beam of visible and near-infrared light. The light is samples come from same person . With this background we
measured with a spectroscope after being scattered by use the terms
the skin. The measurements are analyzed, and a distinct a) FMR (false match rate) as the proportion of the
optical pattern is extracted. time that a biometric template B’ matches B when β≠β’
e) Gait recognition FMR(T)=1-S=t∫∞pn(s)ds where
Recognizing individuals by their distinctive walk, T is the threshold value of s and pn(s) is the non match
involves capturing a sequence of images to derive and distribution.
analyze motion characteristics. A person's gait can be b) FNMR(false non match rate) as the proportion of
hard to disguise because a person's musculature time that a biometric template B’ does not match B when
essentially limits the variation of motion, and measuring β’=β.
it requires no contact with the person. However, gait can A matcher operating at a high threshold T has a low
be disguised if the individual, for example, is wearing FMR and high FNMR;a low threshold means high FMR and
loose fitting clothes. Preliminary results have confirmed low FNMR.A high False Match / False Accept means
its potential, but further development is necessary before security breaches whereas a high False Non Match /False
its performance, limitations, and advantages can be fully Reject means inconvenience. Hence the threshold T
assessed. should be selected such that
f) Ear shape recognition the biometric system operates in an optimal fashion.
Ear shape recognition is still a research topic. It is 3.1.2 ROC CURVE
based on the distinctive shape of each person's ears and The error rates can be plotted against each other as a two
the structure of the largely cartilaginous, projecting dimensional curve called the Receiver Operating
portion of the outer ear. Although ear biometrics appears Characteristic curve.
ROC(T)=(FMR(T),FNMR(T));

ISSN : 0975-3397 2405

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

(1,0) as t-∞ ; (0,1) as t∞ from the family FMR(T)×ρi +FNMR(T)×ρg. = K with K
being the lowest value for such an intersection .

Fig 5. Probability density vs match score curve [1]

Fig 7. Minimum expected error is not at the same operating point as the
equal error rate [1]

d) Cost Functions :
Calculates cost of matcher’s decision by associating a cost
with each of the errors .
Cost = CFM × FMR(T) × ρi + CFNM × FNMR(T) × ρg
Cost functions are used in NIST speaker identification
evaluation system by putting dolor cost for CFM and CFNM.

Fig 6 FMR vs. FNMR curve [1] 3.2 PERFORMANCE METRICS FOR
IDENTIFICATION SYSTEMS
A point on the ROC curve can be chosen as the 3.2.1 THRESHOLD BASED IDENTIFICATION
threshold and this is called the operating point which
specifies the acceptable FMR and FNMR for the biometric Threshold based identification system compares the query
system.ROC curve can be used for rating biometric system biometric B’ to each of the identities in the database , as
based on following matrices. though that identity had been claimed by the subject in a
a) Equal error rate : Equal error point is the operating verification system. This is typically done by computing a
point where FMR = FNMR.For example in figure 6 EERa < score s(B’, Bi) , i = 1,2,3,…m for each enrolled template and
EERb making ‘a’ a better system than ‘b’. However since considering all those with scores exceeding some threshold t0
this metric is based on a narrow range of operating points, , as match. The complete list of all matching identities is
FMR €[ EERa , EERb] or FNMR €[ EERa , EERb] ,this forms returned. Following situations are possible :
an unreliable measure of system accuracy. a) An ideal system which will return m “NO”
b) d-prime: Another way of judging the quality of a answers when the test subject is not enrolled or a single
matcher is to measure how well the non match score “YES” for the correct enrollee when the subject is in the
probability density ρn(s) and the match score probability database.
density ρm(s) are separated. b) More than one candidate may exceed the threshold
d’ = µm-µn /(√(σ 2m+ σ2n)) where , giving an ambiguous candidate list and no definite
µm and σm are the mean and variance of matched scores of identification .
genuine users. c) A single false match may exceed the threshold
µn and σn are the mean and variance of scores of giving a misidentification .
mismatching fingerprints. d) No candidate exceeds the threshold even though
c) Expected overall error: Equal error rate and d- the text subject is enrolled giving a false reject.
prime treat FM and FNM errors as equally likely and of
equal importance which is typically not the case for a 3.2.1.1 FALSE ACCEPT RATE & FALSE REJECT
biometric authentication application. Expected overall error RATE
takes into account the possibility of different FM and FNM
and is given as An imposter is declared as falsely accepted if one or more
E(T) = FMR(T)×ρi +FNMR(T)×ρg. where scores for incorrect candidates exceeds the threshold .The
T = threshold chance of correctly rejecting an imposter is
ρ i= probability of a random user being an imposter. prob (correct reject) = i=1∏m(1-FARi).
ρg. = probability of a random user being genuine. FARi = separately measurable FAR for each identity IDi in
The minimum overall error of a matcher is defined as Emin database M.
and is the point where ROC intersects a diagonal line Ld

ISSN : 0975-3397 2406

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

Simplifying FARi = FAR(the overall system performance candidate list should be empty . In practice , by design , the
parameter) rank engine is still required to return a list of length K.
Prob(correct reject) = (1-FAR) A rank based system can only make a misidentification i.e.
Prob of a false accept FAR(m) = 1- prob(correct reject) = 1- when the true identity is ranked lower than one or more of
(1-FAR)m. the other enrolled candidates. In practice since it is not really
For small FAR , (1-FAR)m ~1-m×FAR Hence FAR(m) = possible to guarantee a closed world; performing
m×FAR identification by always taking the highest ranked candidate
Prob (correct identification) = 1-FRR is not ideal for applications where security is very important.
3.2.2.2 RANK PROBABILITY MASS FUNCTION (RPM)
3.2.2 RANK BASED IDENTIFICATION
:
A rank based identification system returns the top ‘K’ The ranking behavior of the identification system can be
most similar enrolled candidates to the presented biometric characterized by the rank probability mass function P(r)
B’ in some order. Some other secondary decision making defined as P(r) = prob(r) , r = 1…m
process for example a secondary matcher or human operator For any input query B’ corresponding to an enrolled identity
decides on the actual strength of similarities. Such rank ,the probability that the correct identity is ranked in position
based identification system can do true identification only if ‘r’ is P(r) . The RPM can be estimated empirically as P`(K) ,
a) The output candidate vector or list is of length one using a database L = {BL , L = 1…L}of test biometrics and
, i.e. , K = 1. counting the
b) Only enrolled users present their biometric to the
system . Otherwise if the query biometric B’ is not in M, the
cases where the rank of correct match is K indicated by LK.
LK = {B’l ; r(BL’,B) = K ie., when BL’= BK}
then P`(K) = |LK |/|L| , K = 1…m

3.2.2.1 CUMULATIVE MATCH CURVE (CMC)

The main issue to be determined for a rank based
identification system is –what is the optimal candidate list or
vector k. In many cases a fixed k is chosen based on application
logistics. For example there might be space on the display screen
to show only 8 mug shots with good resolution or there might
only be enough human labor available to examine two alternate
fingerprints for every passenger screened. To help choose k we
use a performance statistic of biometric search engines called the
cumulative match curve (CMC).CMC is cumulative sum of
Fig 8. CMC curve [1]
RPM CMC(K) = ∑ki=1 P`(i) , K = 1…m The expected rank
The CMC curve helps us choose k so that the probability that the
E(r) = r` = ∑mk=1 KP`(K) helps evaluate the
correct answer is contained in the shortlist is greater than some
performance of rank based identification system. K should be >
specified performance goal. The CMC increases with k and when k =
E(r) to give a good chance of finding the correct answer in the
m (the whole database ) this probability is one .
candidate list .

4. COMPARISON TABLE SUMMARIZING VARIOUS BIOMETRIC SYSTEMS

Transaction Time
Errors Possible

False Negative
False Positive

Acceptance[4]

Template Size

No of vendors
Identification

Reliability[3]

Intrusiveness
Accuracy[4]
Verification

Stability[3]
Security[3]

Stds Exist?
Error Rate
Biometric

Low Cost
System
S.no

1 Finger √ √ 4 3 1 in Dryness,di Ext diff Ext 3 3 2 Some √ 80- 2- Y >2

Print 500 + rt,age difficult what 1000b 10sec 5
ytes
2 Facial √ × 3 2 No Lighting,a Diffcult Easy 2 2 2 Non √ 1-4KB 2- ? 7
Recog Data ge,glasses, 10sec
nition hair
3 Hand √ × 3 2 1 in Hand Very Medium 2 2 2 Non × 9bytes 3- ? 1
Geom 500 injury, age difficult 10sec

ISSN : 0975-3397 2407

 Jammi Ashok et. al. / (IJCSE) International Journal on Computer Science and Engineering
Vol. 02, No. 07, 2010, 2402-2408

etry
4 Iris √ √ 4 3 1 In Poor Very Very 3 3 2 Non × 512byt 10sec ? 1
Scan 131,00 lighting difficult difficult es
0
5 Retina √ √ 4 3 1 in Glasses Ext Ext 3 3 2 Very × Mode ? 3
l Scan 10,000 difficult difficult rate
,000+
6 Vein √ × 2 3 No ----- Very Medium 3 3 2 Non × Mode ? 1
Scan data difficult rate
7 Facial √ × 1 2 Very Medium 3 3 2 Non × Mode ? 1
Therm difficult rate
o -----
gram

5. CONCLUSION 7. AUTHORS BIOGRAPHY

The conservative authentication methods are based
either on physical possessions of a person like, ID cards , Prof J.Ashok is currently working as
Keys, etc. OR the person’s secret knowledge like passwords/ Professor and Head of Information
PIN . All these can however be faked with some effort. Technology at Geethanjali College of
Also, carrying them may be cumbersome and unsuitable in Engg. & Technology, Hyderabad, A.P,
certain areas of high security / risk zones like military INDIA. He has received his B.E.
operations. Biometrics emerged as an effective panacea to Degree from Electronics and
this serious shortcoming as biological features cannot be Communication Engineering from
faked easily and there is no overhead to be carried or Osmania University and M.E. with
remembered. specialization in Computer Technology
A new age of Identity Authentication has thus arrived. from SRTMU, Nanded, INDIA. His
Plastic ID cards, paper forms, multiple passwords and pins main research interest includes neural networks, data retrieval
are slowly but surely making way for the Biometrics. process and Artificial Intelligence. He has been involved in the
Research is presently on to combine Biometrics with the organization of a number of conferences and workshops. He
emerging revolutionary technology, the nanotechnology the has been published more than 30 papers in national and
science of altering materials at molecular levels. At this International journals and conferences. He is currently doing
stage, the ideas are still not fully formed, and products are his Ph.D from Anna University and is at the end of
emerging but they will undergo radical changes as submission.
nanotechnology starts evolving. Both biometrics and
Mr. VAKA SHIVASHANKAR is
nanotechnology combined could create products too small to
currently working as Asst.professor in
be viewed by the human eye. A human being’s entire life
Kakatiya College of computer
history could be stored in a chip smaller than a grain of sand.
sciences,Warangal,A.P.,INDIA.
This type technology is under research in the military.
He completed his BCA and MCA from
Kakatiya University,
6. REFERENCES
Warangal,A.P.INDIA. HE is doing
[1] Rudd M.Bolle,Jonathan H.Connell,Sharath Pankanti , Nalini K . M.Tech (CSE) from JNTU. His main
Ratha , Andrew W . Senior, Guide to biometrics , Springer research interest includes Artificial
Publication (2003). Intelligence, Computer Networks, and Data Mining.
[2] Davide Maltoni, Anil K. Jain ,Handbook of fingerprint
recognition,Springer publication (2002). Mr.P.V.G.S.Mudiraj is currently
[3] Dr. Sujoy Bhattacharya,Consultant (TATA Consultancy Services working as Associate Professor in the
Ltd.), Digital Image Processing Concepts and Trends Lecture notes
[4] J.G.Daugman,High confidence visual recognition of persons by a
department of MCA at Adams Engg.
test of statistical impedance ,IEEE Transactions on pattern analysis College ,Paloncha, A.P, INDIA. He
and machine intelligence,15(11):1148-1161,Nov 1993 received his M.Tech with Specialization
[5] http://www.findbiometrics.com in Information Technology from IASE
[6] http://www.biometricsinfo.org
[7] http://www.biometrics.gov/
University, Rajasthan He received his
[8] Biometrics: Retinal Scanning Amy Zalman MCA from Osmania University,
[9] http://www.tiresias.org/ Hyderabad INDIA. His main research
[10] http://www.biovericom.com/biotech/retina.html interest includes Information Security, data mining and data
[11] www.howstuffworks.com
[12] http://www.globalsecurity.org/security/systems/biometrics-
ware housing and Bio Informatics. Currently he is doing his
emerging.htm Ph.D(CSE) from Rayalaseema University
[13] www.authentec.com

ISSN : 0975-3397 2408