ijest10-02-07-24

8/2/2019 IJEST10-02-07-24

1/24

G.Prasanna Lakshmi / International Journal of Engineering Science and Technology

Vol. 2(7), 2010, 3054-3077

FINGERPRINT IDENTIFICATION SYSTEM combined with

CRYPTOGRAPHY for Authentication.

G.Prasanna Lakshmi*

Research Scholar ,Computer Science, GITAM University, Opposite Rushikonda Beach,

Visakhapatnam, Andhra Pradesh-530045, India

[email protected] Phone No :9324230128

Abstract:Biometrics technology, which uses physical or behavioral characteristics to identify users, has come to attract

increased attention as a means of reliable personal authentication that helps the identity of an actual user.

Among various modalities of Biometrics, Fingerprints are known to have the longest history of actual use in law

enforcement applications with proven performance. This project surveys the state of the art in fingerprint

identification technology. In this project, a design schema of a security authentication system combined with

fingerprint identification and public key cryptography is explored, and its specific security mechanism is

discussed in detail. In our schema, fingerprint is added into user's private key and served a security parameter,such that users secret key is separated into secret key parameters and fingerprint, by secret splitting mechanism,

which makes the secret key to be bounded with user's information. This will increase the security of secret key

ultimately. In such an authentication system, the diplex authentication technologies --- fingerprint and smart

card --- are adopted, and the user fingerprint neednt to be transmitted during the authentication process, which

can protect user's privacy effectively.

Keywords: Biometric Cryptography 1; Minutiae2; Encryption3.

1. Introduction.

Verifying the identity of an individual can be done through three main methods; what an individual has, what an

individual knows or owns, and what an individual is. The first method is typically achieved through the use of a

token, such as an identification card, badge, magnetic stripe, or Radio Frequency Identification (RFID) tag. The

second method can be achieved through the use of a password, or personal identification number (PIN), and the

third method can be accomplished through what an individual is, more formally known as biometrictechnologies. Similar to the first two authentication methods, biometric systems too contain vulnerabilities and

are susceptible to attack. Some of these vulnerabilities are similar or even overlapping across all three

authentication mechanisms. However, attacks specific to biometric systems focus on liveness detection of a

human i.e. is this finger from a live sample, or a gelatin sample. There have been various documented attacks in

the literature which examine the attack on the sensor . While understanding and preventing attacks on thesensorare interesting research topics in need of investigation, this project examines the global and local features of a

live sample compared to that of a gelatin finger from the same user after acquisition on a commercially available

biometric fingerprint sensor.

2. Biometric System Vulnerabilities

All security measures, including mechanisms for authenticating identity, have ways of being circumvented.

Certainly the processes in working around these measures vary in difficulty based on effort and resources

needed to carry out the deceptive act. Authentication mechanisms based on secrets are particularly vulnerable to"guessing" attacks. Token mechanisms that rely on the possession of an object, most notably a card or badge

technology are most vulnerable to theft or falsified reproduction. Biometric technologies closely tie the

authenticator to individual identity of the user through the use of physiological or behavioral characteristics.

While this property is an added advantage over the previous two authentication mechanisms mentioned; it

places a great emphasis on validating the integrity of the biometric sample acquired and transferred in the

biometric system. Ratha, N., et al. provided a model identifying vulnerabilities in biometric systems [3]. An

.

ISSN: 0975-5462 3054

8/2/2019 IJEST10-02-07-24

2/24


Vol. 2(7), 2010, 3054-3077

example of the threat model is shown below in Figure 1, and builds on the general biometric model outlined in

Mansfield and Wayman [4].

Figure 1: Biometric Threat Model

2.1 FINGERPRINT FEATURES

In fingerprint features are classified into three classes. Level 1 features show macro details of the ridge flow

shape, Level 2 features (minutiae point) are discriminative enough for recognition, and Level 3 features (pores)

complement the uniqueness of Level 2 features.

Global Ridge Pattern

A fingerprint is a pattern of alternating convex skin called ridges and concave skin called valleys with a spiral-

curve-like line shape (Figure 1.1.5). There are two types of ridge flows: the pseudo-parallel ridge flows and

high-curvature ridge flows which are located around the core point and/or delta point(s). This representation

relies on the ridge structure, global landmarks and ridge pattern characteristics. The commonly used global

fingerprint features are:

singular points discontinuities in the orientation field. There are two types of singular points. A core is the

uppermost of the innermost curving ridge.

Figure: Global Ridge Patterns

ISSN: 0975-5462 3055

8/2/2019 IJEST10-02-07-24

3/24


Vol. 2(7), 2010, 3054-3077

And a delta point is the junction point where three ridge flows meet. They are usually used for fingerprint

registration, fingerprint classification.

Ridge orientation map local direction of the ridge-valley structure. It is commonly utilized for

classification, image enhancement, and minutia feature verification and filtering.

Ridge frequency map the reciprocal of the ridge distance in the direction per-pendicular to local ridge

orientation. It is formally defined in [32] and is extensively utilized for contextual filtering of fingerprint

images.This representation is sensitive to the quality of the fingerprint images [36]. However, the discriminative

abilities of this representation are limited due to absence of singular points.

(a) A ridge ending minutia: (x,y) are the minutia coordinates; is

the minutias orientation

(b) A ridge bifurcation minutia: (x,y) are the minutia coordinates; is the minutias orientation.

Figure: Minutia ridge patterns

Figure: Minutiae relation model.

ISSN: 0975-5462 3056

8/2/2019 IJEST10-02-07-24

4/24


Vol. 2(7), 2010, 3054-3077

Block Diagrams

Figure 2 : Fingerprint Enrollment & Authentication system

2.1.1 ACTUAL SYSTEM

Fingerprints are the most widely used biometric feature for person identification and verification in the field of

biometric identification. This module re presents the implementation of a minutiae based approach to fingerprint

ISSN: 0975-5462 3057

8/2/2019 IJEST10-02-07-24

5/24


Vol. 2(7), 2010, 3054-3077

identification and verification with cryptography. An ideal scheme for combining physical identity (biometric

features) with logical identity (key)

Biometric features are permanently associated with the user and can be used for identification.

Protection of biometric data itself is a privacy issue Biometrics cannot be revoked

Biometric cryptography:

Combining Biometrics and Cryptography Use biometrics to generate cryptographic keys Successful biometric verification generates correct key

The major steps involved in FINGERPRINT RECOGNITION:

1. Fingerprint database

2. Fingerprint features database3. Enrollment Module

4. Authentication module.

Figure 3: Steps in Fingerprint classification & recognition

Figure: data collection is done using sensor Figure: preprocessing block output

ISSN: 0975-5462 3058

8/2/2019 IJEST10-02-07-24

6/24


Vol. 2(7), 2010, 3054-3077

Figure: after wavelet transformation Image Figure: Feature extraction

is converted into LL, LH,HL & HH

frequency components

3.FEATURE SELECTION

DISCRETE WAVELET TRANSFORM :

The main idea is the same as it is in the CWT. A time-scale representation of a digital signal is obtained using

digital filtering techniques. Recall that the CWT is a correlation between a wavelet at different scales and the

signal with the scale (or the frequency) being used as a measure of similarity. The continuous wavelet transform

was computed by changing the scale of the analysis window, shifting the window in time, multiplying by the

signal, and integrating over all times. In the discrete case, filters of different cutoff frequencies are used to

analyze the signal at different scales. The signal is passed through a series of high pass filters to analyze the high

frequencies, and it is passed through a series of low pass filters to analyze the low frequencies.

ThThis procedure can mathematically be expressed asis procedure can mathematically be expressed as

:

Having said that, we now look how the DWT is actually computed: The DWT analyzes the signal at different

frequency bands with different resolutions by decomposing the signal into a coarse approximation and detail

information. DWT employs two sets of functions, called scaling functions and wavelet functions, which are

associated with low pass and high pass filters, respectively. The decomposition of the signal into different

frequency bands is simply obtained by successive high pass and low pass filtering of the time domain signal.

The original signal x[n] is first passed through a half band high pass filter g[n] and a low pass filter h[n]. After

the filtering, half of the samples can be eliminated according to the Nyquists rule, since the signal now has a

ISSN: 0975-5462 3059

8/2/2019 IJEST10-02-07-24

7/24


Vol. 2(7), 2010, 3054-3077

highest frequency of /2 radians instead of . The signal can therefore be subsampled by 2, simply by

discarding every other sample. This constitutes one level of decomposition and can mathematically be expressed

as follows

g

Figure: Discrete Wavelet Transform Tree

Figure: Divisions of image after D.W.T.

3.1 FEATURE EXTRACTION

Its combining process of cropping & centralizing. Before going to perform these, the image has to enhanced.

ISSN: 0975-5462 3060

8/2/2019 IJEST10-02-07-24

8/24


Vol. 2(7), 2010, 3054-3077

Distortion of fingerprint is more serious issue:

1 A fingerprint image is a 2D image of a 3D finger

2 2D image is affected by pressure, scratches, sweat, alignment and

position of finger.

3 Applying weiner filter is more suitable.

Figure: Cropping and Centralization

Features

Centralizing Cropping

Weiner filter

.Center

ISSN: 0975-5462 3061

8/2/2019 IJEST10-02-07-24

9/24


Vol. 2(7), 2010, 3054-3077

3.1.1 WIENER FILTER

Wiener Filter Perform 2-D adaptive noise-removal filtering.

Wiener low-pass filters an intensity image that has been degraded by constant power additive noise. Wiener

filter uses a pixel-wise adaptive Wiener method based on statistics estimated from a local neighborhood of

each pixel.

In the Digital Filters and Z Transforms chapter we introduced inverse filters as a way of undoing some

instrumental effect to determine the "true" signal. Later in that chapter, we saw that if we have a filter which has

a large number of terms in it:

then we can do "another implementation" of the same filter in terms of its inverse that may have fewer

significant terms in it:

4. PRIVACY & SECURITY ISSUES IN BIOMETRIC

SYSTEMS:

Figure: Privacy & Security issues in Fingerprint identification system.

ISSN: 0975-5462 3062

8/2/2019 IJEST10-02-07-24

10/24


Vol. 2(7), 2010, 3054-3077

4.1 Conventional Cryptography

In conventional cryptography, also called secret-key or symmetric-key encryption, one key is used both for

encryption and decryption. The Data Encryption Standard (DES) is an example of a conventional cryptosystem

that is widely employed by the Federal Government. Figure 1-1 is an illustration of the conventional encryption

process.

Figure: conventional encryption

Figure: conventional decryption

ISSN: 0975-5462 3063

8/2/2019 IJEST10-02-07-24

11/24


Vol. 2(7), 2010, 3054-3077

4.1.2 RSA ALGORITHM

Public Key Cryptography:

One of the biggest problems in cryptography is the distribution of keys. Suppose you live in the United States

and want to pass information secretly to your friend in Europe. If you truly want to keep the information secret,

you need to agree on some sort of key that you and he can use to encode/decode messages. But you don't want

to keep using the same key, or you will make it easier and easier for others to crack your cipher. But it's also a

pain to get keys to your friend. If you mail them, they might be stolen. If you send them cryptographically, and

someone has broken your code, that person will also have the next key. If you have to go to Europe regularly to

hand-deliver the next key, that is also expensive. If you hire some courier to deliver the new key, you have to

trust the courier, etcetera.

Certification

There is, of course, a problem with the scheme above. Since the public keys are really public, anyone can send a

message to you. So your enemy can pretend to be your friend and send you a message just like your friend as

they both have access to the public key. Your enemy's information can completely mislead you. So how can you

be certain that a message that says it is from your friend is really from your friend? Here is one way to do it,

assuming that you both have the public and private keys Ea, Eb, Da, and Db as discussed in the previous

section. Suppose I wish to send my friend a message that only he can read, but in such a way that he is certain

that the message is from me. Here's how to do it. I will take my name, and pretend that it is an encoded message,

and decode it using Da. I am the only person who can do this, since I am the only person who knows Da. Then I

include that text in the real message I wish to send, and I encode the whole message using Eb, which only my

friend knows how to decode. When he receives it, he will decode it using Db, and he will have a message with

an additional piece of what looks to him like junk characters. The junk characters are what I got by decoding my

name. So he simply encodes the junk using my public key Ea and makes certain that it is my name. Since I am

the only one who knows how to make text that will encode to my name, he knows the message is from me. You

can encode any text for certification, and in fact, you should probably change it with each

message, but it's easy to do.

RSA Encryption and Decryption:

One commonly used cipher form is called RSA Encryption, where RSA are the initials of the three creators:

Rivest, Shamir, and Adleman. It is based on the following idea: It is very simply to multiply numbers together,

especially with computers. But it can be very difficult to factor numbers. In practical applications, it is common

to choose a small public exponent for the public key. In fact, entire groups of users can use the same public

exponent, each with a different modulus. (There are some restrictions on the prime factors of the modulus when

the public exponent is fixed.) This makes encryption faster than decryption and verification faster than signing.

ISSN: 0975-5462 3064

8/2/2019 IJEST10-02-07-24

12/24


Vol. 2(7), 2010, 3054-3077

With the typical modular exponentiation algorithms used to implement the RSA algorithm, public key

operations take O(k2) steps, private key operations take O(k3) steps, and key generation takes O(k4) steps,

where k is the number of bits in the modulus. Fast multiplication techniques, such as methods based on the

Fast Fourier Transform (FFT), require asymptotically fewer steps. In practice, however, they are not as common

due to their greater software complexity and the fact that they may actually be slower for typical key sizes.

Encryption Operation of RSA

Figure: encryption operation of RSA

Entity A wants to send secret message(cipher text) to B

Uses Public key of B and Encrypt operation to generate Ciphertext

Sends Ciphertext to Entity B

Entity B wants to read the Message(M) sent by A

Receives Ciphertext from A Uses its private key D and Decrypt operation to get Message(M)

Authentication Operation of RSA

Figure: authentication operation of RSA

A BCiphertext

Public Key of

B

Message(M)

Encrypt

operation

Ciphertext

Private Key D

CiphertextDecrypt

operation

Message(M)

A

BSignature (S), Message (M)Private key D

Message(M)

Signing

operation

Signature(S)

Public key Eof A

Signature(S)

Verifying

operationMessage(M1)

If M equals M1

then identity ofEntity A isproved

ISSN: 0975-5462 3065

8/2/2019 IJEST10-02-07-24

13/24


Vol. 2(7), 2010, 3054-3077

Entity A wants to prove its identity to B

Signing operation to generate Signature (S)

sends Message(M), Signature(S) to Entity B

Entity B wants to verify identity of A

Receives Signature(S) and Message(M) from A

Verifying operation to generate Message(M1) from S

Compares M1 and M to verify identity of A

5 .ALGORITHM

The algorithm can be given below.

1. Find P and Q, two large (e.g., 1024-bit) prime numbers.

2. Choose E such that E is greater than 1, E is less than PQ, also E and (P-1)(Q-1) are relatively prime,

which means they have no prime factors in common. E does not have to be prime, but it must be odd.

(P-1)(Q-1) can't be prime because it's an even number.

3. Compute D such that (DE - 1) is evenly divisible by (P-1)(Q-1). Mathematicians write this as

DE = 1 (mod (P-1)(Q-1)), and they call D the multiplicative inverse of E. This is easy to do simply find

an integer X which causes D = (X(P-1)(Q-1) + 1)/E to be an integer, then use that value of D.

4. The encryption function is C = (T^E) mod PQ, where C is the cipher text (a positive integer), T is the

plaintext (a positive integer), and ^ indicates exponentiation. The message being encrypted, T, must be

less than the modulus, PQ.

5. The decryption function is T = (C^D) mod PQ, where C is the cipher text (a positive integer), T is the

plaintext (a positive integer), and ^ indicates exponentiation.

Your public key is the pair (PQ, E). Your private key is the number D (reveal it to no one). The product PQ is

the modulus (often called N in the literature). E is the public exponent. D is the secret exponent.

You can publish your public key freely, because there are no known easy methods of calculating D, P, or Q

given only (PQ, E) (your public key). If P and Q are each 1024 bits long, the sun will burn out before the most

powerful computers presently in existence can factor your modulus into P and Q.

5.1 RSA-CRT key generation

1. Let p and q be very be two very large primes of nearly the same size such that gcd

(p-1, q-1) = 2.

2. Compute N = p*q. and Phi=(p-1)*(q-1);

3. Pick two random integers dp and dq such that gcd (dp, p-1)=1, gcd (dq, q-1)=1 and

dp==dq mod 2.

ISSN: 0975-5462 3066

8/2/2019 IJEST10-02-07-24

14/24


Vol. 2(7), 2010, 3054-3077

4. Find a d such that d==dp mod p-1 and d==dq mod q-1.

5. Compute e=d-1 (mod Phi ).

The public key is and the private key is . Since gcd (dp, p-1)=1 and d==dp mod p-1, we

have gcd (d, p-1)=1. Similarly, gcd (d, q-1)=1. Hence gcd (d, phi (N))=1 and by step 5, e can be computed.

To apply the Chinese Remainder Theorem in step 4, the respective moduli have to be relatively prime in pairs

for a solution to necessarily exist. We observe that p-1 and q-1 are even and that we cannot directly apply the

Chinese Remainder Theorem. However, gcd ((p- 1)/2, (q-1)/2)=1. Since gcd (dp, p-1)=1 and gcd (dq, q-1)=1,

essentially dp, dq are odd integers and dp-1, dq-1 are even integers. We have gcd (d, p-1)=1, which implies that

d is odd and d-1 is even.

To find a solution to

d==dp mod p-1,

d==dq mod q-1.

We find a solution to

d-1==dp 1 mod p-1,

d-1==dq 1 mod q-1.

By applying the cancellation law and taking the common factor 2 out, we have

x=d== (d-1)/2==(dp 1)/2 mod( p-1)/2,

x=d==(d-1)/2==(dq 1)/2 mod( q-1)/2.

Using Chinese Remainder Theorem we find d such that d = (2*d) +1.

5.1.2 RSA-CRT Decryption

Since RSA-CRT encryption is same as that of the standard RSA encryption procedure, we now turn our

attention to RSA-CRT decryption. Let M be the plaintext and C the cipher text.

Theorem If C is not divisible by p and dp==d mod p-1, then Cdp==Cd (mod p).

For decryption we find

1. Mp=Cdp(mod p)= Cd(mod p) and Mq=Cdq(mod q)= Cd(mod q).

2. Then using Chinese Remainder Theorem, we find a solution for

M=Mp(mod p)= Cd(mod p),

M=Mq=Cdq(mod q)= Cd(mod q).

We now illustrate the scheme using an over simplified example. Choose p = 7, q = 11, gcd

(p-1, q-1) = 2, N = p*q = 7*11 = 77, phi (N) = (p-1)*(q-1) = 6*10 = 60.

Let dp = 5, gcd (dp , p-1) = gcd (5,6) = 1.

dq = 3, gcd (dq , q-1) = gcd (3,10) = 1.

ISSN: 0975-5462 3067

8/2/2019 IJEST10-02-07-24

15/24


Vol. 2(7), 2010, 3054-3077

We are to find d such that

d==5 mod 6,

d==3 mod 10.

We cannot apply the Chinese Remainder theorem since gcd (6,10) . 1, hence we convert the system of

congruences in such a manner that the cancellation law can be applied

Therefore, we have

d-1==5-1 mod 6,

d-1==3-1 mod 10.

On applying the cancellation law,

(d-1)/2==(5-1)/2 mod (6/2),

(d-1)/2==(3-1)/2 mod (10/2).

x = d= (d-1)/2== 2 mod 3,

x = d= (d-1)/2== 1 mod 5.

Solving using Chinese Remainder Theorem,

M = 3*5 = 15, M1 =15/3 = 5, M2 = 15/5=3.

5*N1==1 mod 3, N1=2,

3*N2==1 mod 5, N2=2.

We have,

d = x = 2*5*2 + 1*3*2 = 26(mod 15) = 11.

Therefore d = 11 and d = (2*d)+1 = (2*11) +1 = 23, d = 23.

Now we find, e such that

e*d==1 mod phi(N),

e*23==1 mod 60, e = 47.

Let the plaintext M=5.

C=547 mod 77 = 3.

For decryption, we find

M = Mp mod p = cd mod p,M = Mq mod q = cd mod q.

Mp = 35 mod 7 = 243 mod 7 = 5,

Mq = 33 mod 11= 27 mod 11 = 5.

Using the Chinese Remainder Theorem,

ISSN: 0975-5462 3068

8/2/2019 IJEST10-02-07-24

16/24


Vol. 2(7), 2010, 3054-3077

M = 7*11 = 77, M1 = 77/7 = 11, M2 = 77/11 = 7.

11*N1==1 mod 7, N1=2,

7*N2==1 mod 11, N2=8.

x = 5*11*2 + 5*7*8 = 390 mod 77 =5.

Thus x = M = 5, as desired. In this specific example (Mp and Mq)=5 is a common solution and it is not

necessary to further apply the Chinese Remainder Theorem.

5.1.3 CRT to Multi-Prime RSA:

Based on the property of the RSA algorithm, the modulus N is the product of large prime numbers. Thus we can

use Chinese Remainder Theorem (CRT) to accelerate the computation. In 2-prime CRT and 2-prime RSA, the

modulus

S1=M D mod p

S2=M D mod q

By applying Fermats theorem, we can obtain

S1=M D 1 mod p

S2=M D 2 mod q

where D1=D mod (p-1) and D2=D mod(q-1). Applying CRT we can compute the results

S as

S = (S1c1q+S2c2p) mod N 2

Where

c1=q-1 mod p

c2=p-1 mod q.

The size of p and q about half of N. Thus the size of the exponents is reduced to half of the original size in 2-

prime RSA. In multi prime CRT and RSA

we have S=MD mod (pqr)

We can obtain

S1=MD1 mod p

S2=MD2 mod q

S3=MD3 mod r.

Where,

D1=Dmod(p-1)

D2=Dmod(q-1)

D3=Dmod(r-1)

ISSN: 0975-5462 3069

8/2/2019 IJEST10-02-07-24

17/24


Vol. 2(7), 2010, 3054-3077

We can apply the CRT to retrive S as S=(S1c1qr+ S2c2pr+ S3c3pq)

Where

C1=(rq)-1 mod p

C2=(pr)-1 mod q

C3=(pq)-1 mod r

Hence the size of the exponents is further reduced to one third the original. Based on above analysis, 1024-bit 2-

prime and multi-prime RSA can be done with 512-bit and 341-bit exponents and modulus respectively.

7 Results and Conclusions

GUI

ISSN: 0975-5462 3070

8/2/2019 IJEST10-02-07-24

18/24


Vol. 2(7), 2010, 3054-3077

ISSN: 0975-5462 3071

8/2/2019 IJEST10-02-07-24

19/24


Vol. 2(7), 2010, 3054-3077

ISSN: 0975-5462 3072

8/2/2019 IJEST10-02-07-24

20/24


Vol. 2(7), 2010, 3054-3077

ISSN: 0975-5462 3073

8/2/2019 IJEST10-02-07-24

21/24


Vol. 2(7), 2010, 3054-3077

ISSN: 0975-5462 3074

8/2/2019 IJEST10-02-07-24

22/24


Vol. 2(7), 2010, 3054-3077

8 APPLICATIONS

Markets for fingerprint technology include entrance control and door-lock applications, fingerprint identification

mouses, fingerprint mobile phones, and many others. The fingerprint markets are classified as

follows:

Figure: market applications of fingerprint identification

As the advanced technology enables even more compact fingerprint sensor size, the range of application is

extended to the mobile market. Considering the growing phase of the present mobile market, its potential is the

greatest of all application markets.

ISSN: 0975-5462 3075

8/2/2019 IJEST10-02-07-24

23/24


Vol. 2(7), 2010, 3054-3077

SYSTEMSECURITY:

Figure: role of bad user

If there is only fingerprint or any of the biometric technology is used, then there may be a chance of a bad user

tampering or corrupting the data. so we are combining the fingerprint technology with cryptography which

results in a totally secured system.

Figure: cryptography in system security

ISSN: 0975-5462 3076

8/2/2019 IJEST10-02-07-24

24/24


Vol. 2(7), 2010, 3054-3077

The problem faced by the single biometric system is solved by combining the system with cryptography which

is yielding many advantages & applications in real time.

References

[1] R. Cappelli, Synthetic Fingerprint Generation,Handbook of Fingerprint Recognition, D. Maltoni, D. Maio, A.K. Jain, and S.

Prabhakar, eds. New York: Springer, 2003.[2] R. Cappelli, A. Erol, D. Maio, and D. Maltoni, Synthetic Fingerprint-Image Generation, Proc. 15th Int'l Conf. Pattern Recognition,pp.

475-478, Sept. 2000.

[3] R. Cappelli, D. Maio, and D. Maltoni, Modelling Plastic Distortion in Fingerprint Images, Proc. Second Int'l Conf. Advances in

Pattern Recognition, pp. 369-376, Mar. 2001.[4] R. Cappelli, D. Maio, and D. Maltoni, Synthetic Fingerprint-Database Generation, Proc. 16th Int'l Conf. Pattern Recognition, vol. 3,

pp. 744-747, Aug. 2002.

[5] Y. Dit-Yan et al., SVC2004: First International Signature Verification Competition, Proc. Int'l Conf. Biometric Authentication, pp. 16-

22, July 2004.

[6] G. Doddington et al., Sheep, Goats, Lambs and Wolves: A Statistical Analysis of Speaker Performance, Proc. Int'l Conf. Language

and Speech Processing, pp. 1351-1354, Nov. 1998.[7] S. Pankanti, N.K. Ratha, and R.M. Bolle, Structure in Errors: A Case Study in Fingerprint Verification, Proc. 16th Int'l Conf. Pattern

Recognition, 2002.

[8] P.J. Phillips, A. Martin, C.L. Wilson, and M. Przybocky, An Introduction to Evaluating Biometric Systems, Computer, vol. 33, no. 2,

Feb. 2000.[9] P.J. Phillips, H. Moon, S.A. Rizvi, and P.J. Rauss, The FERET Evaluation Methodology for Face-Recognition Algorithms,IEEE

Trans. Pattern Analysis and Machine Intelligence, vol. 22, no. 10, pp. 1090-1104, Oct. 2000.

ijest10-02-07-24

Documents