ArticlePDF Available

Recent Developments on Applying Biometrics in Cryptography

January 2010
Journal of Applied Security Research 5(1):107-137

January 2010
5(1):107-137

DOI:10.1080/19361610903176328

Authors:

Biometrics is an identification technology widely used in physical access control and forensics. Cryptography is the science for information security. People generally agreed that biometric information needs to be protected with cryptography. Many researchers also believe that biometrics can enhance the security of cryptography. In this article I have surveyed recent developments on applying biometrics in cryptography, focusing on binding and generating a cryptographic key. The security of these proposals is briefly analyzed. Additionally I have reviewed some other developments, such as matching without removing chaff points, encrypted templates matching, symmetric hashing, and biometric random number generation.

Some Implementations of Biometric Key Generation and Binding

…

Content may be subject to copyright.

Content uploaded by Qinghai Gao

Content may be subject to copyright.

Journal of Applied Security Research, 5:107–137, 2009

ISSN: 1936-1610 print / 1936-1629 online

DOI: 10.1080/19361610903176328

TECHNOLOGY

Recent Developments on Applying Biometrics

in Cryptography

QINGHAI GAO, PhD

Farmingdale State College, A Campus of the State University of New York,

Farmingdale, New York, USA

Biometrics is an identiﬁcation technology widely used in physical

access control and forensics. Cryptography is the science for infor-

mation security. People generally agreed that biometric informa-

tion needs to be protected with cryptography. Many researchers also

believe that biometrics can enhance the security of cryptography. In

this article I have surveyed recent developments on applying biomet-

rics in cryptography, focusing on binding and generating a crypto-

graphic key. The security of these proposals is brieﬂy analyzed. Ad-

ditionally I have reviewed some other developments, such as match-

ing without removing chaff points, encrypted templates matching,

symmetric hashing, and biometric random number generation.

KEYWORDS Biometrics, cryptography, key, generation, binding,

template, security

INTRODUCTION

Biometrics is deﬁned as the identiﬁcation of an individual based on phys-

iological and behavioral characteristics. Main physiological characteristics

include face (2D/3D facial images, facial IR thermogram), hand (ﬁngerprint,

hand geometry, palmprint, hand IR thermogram), eye (iris and retina), ear,

skin, odor, dental, and DNA. Main behavioral characteristics include voice,

gait, keystroke, signature, mouse movement, and pulse. One or more of

the aforementioned biometrics can be combined in a system to improve the

The author thanks Professor John Kostanoski, Chair, Department of Security Systems,

Farmingdale State College, SUNY for his assistance in the preparation of this article.

Address correspondence to Qinghai Gao, PhD, Assistant Professor, Department of Secu-

rity Systems, Farmingdale State College, A Campus of the State University of New York, 2350

Broadhollow Road, Farmingdale, NY 11735. E-mail: gaoqj@farmingdale.edu

107

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

108 Q. Gao

accuracy of recognition. In addition, Jain et al. (2004a, 2004b) also proposed

to use soft biometric traits like gender, age, height, weight, ethnicity, and

eye color to assist in identiﬁcation.

Overall biometric systems are designed to solve a matching problem

through the live measurements of biometrics. It operates with two steps

(Maltoni et al., 2003; Jain et al., 1997; Jain et al., 2000). First, a person must

Second, the person must provide the same biometric for new measurements.

The output of the new measurements will be processed with the same

algorithms as those used at registration and then compared to the stored

template. If the similarity is greater than a system-deﬁned threshold, the

veriﬁcation is successful; otherwise it will be considered unsuccessful.

One popular application of biometrics is password management (Klein,

1990; Armstrong, 2003; Feldmeier & Karn, 1990; O’Gorman, 2003; Skaff,

2007). As we know, text-based passwords are the most common way for

authenticating users. Password-based online authentication also works with

two stages. During registration, a user selects a password and then the hash

of the password will be stored in a database. During authentication, the

newly entered password will be hashed locally, and the new hash will be

sent over the Internet to the server for comparison with the stored hash.

Only if the two hashes match exactly is authentication successful; otherwise

it is unsuccessful. In general, a unique username is required for each account

because the uniqueness of a username guarantees the hash matching is a

1-to-1 veriﬁcation process.

The advantage of password-based authentication is that passwords are

short text strings without containing any non-deterministic bit, thus it can

be memorized and hashed in real-time with a one-way hash function like

SHA-1 (Secure Hash Algorithm).

However, there are a few problems with this scheme. Many computer

users have too many passwords and tend to choose easy-to-remember pass-

words that can be hacked easily. Non-repudiation is another problem be-

cause a password can be transferable easily. Therefore, people have been

looking for better ways to achieve high security.

One proposed solution to the problem is to use biometric authentication

because biometrics are physically with us and cannot be transferred. Unlike

a password it is unnecessary for a user to know the detailed information of

the chosen biometric. All that needs to be remembered is to use the same

registered biometric.

Another possible usage of biometrics is to improve the security of cryp-

tography. Typically, a cryptographic key is generated mathematically and

then assigned arbitrarily to a user. This approach has two problems. First is

the repudiation problem due to the lack of direct physical connection be-

tween the key and its owner. Secondly, the key has to be stored somewhere

because it is too long to be memorized. An easy-to-remember password is

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 109

then utilized to access the stored key. Therefore the security of the system

is just as strong (or weak) as the password. By introducing biometrics into

cryptography the two problems potentially could be solved, for example,

by biometrically controlling the access of the stored key or by generating a

strong cryptographic key directly from biometrics.

Since the 1990s the mainstream research in biometrics has been focused

on image processing and pattern recognition, trying to improve the accu-

racy of biometric authentication systems. Only in the past few years has

research that focused on introducing biometrics into cryptography attracted

considerable interest. In this article I will attempt to summarize the recent

developments in the area and propose new research objectives.

What follows next is a discussion on the methods to solve the non-

reproducible problem of biometrics; a review of different biometrics that

have been proposed for cryptographic key generation and binding; a review

of the necessity of protecting biometric templates; a discussion on biometric

random number generator; and a summary and prediction of possible future

research directions.

METHOD TO COPE WITH NON-REPRODUCIBILITY

There is inherent fuzziness with biometric measurements due to changes

in physical and environmental conditions. Non-reproducibility lies at the

heart for any biometric application, including biometric key generation and

binding. To date it is still a very challenging problem to achieve security

with noisy data (Tuyls et al., 2007). In this section I attempt to summarize

the common methods found in the literature, without listing any particular

reference.

Averaging/Training

During registration, a number of biometric samples with some variations

were obtained, transformed, and then averaged to get a generic represen-

tation of the biometric. During authentication, the same biometric will be

measured and compared with the stored representation. The mean and stan-

dard deviation of the samples are often needed.

The biometric samples collected during registration can also be applied

to train a mathematical model, such as Hidden Markov Model (HMM), Sup-

port Vector Machine (SVM), and some clustering techniques. The parameters

obtained at the end of training will be used during veriﬁcation.

Quantization/Tessellation/Discretization

The image of a biometric will be quantized into a number of small units.

The biometric information inside each unit will be assumed to locate at the

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

110 Q. Gao

center of the unit. This quantization (also called tessellation or discretization)

process not only can remove large amounts of noises, but also can reduce

the overall dimensions of the biometric images.

Majority Voting

For a number of measurements of a biometric, each will be quantized and

binarized into a ﬁxed-length string. For every bit position of the binary

strings, majority voting will be used to determine the value, 0 or 1.

Error Correction Coding and Helper Data

Error correction coding is often used for noisy data. Two common choices

are Reed-Solomon (RS) coding and Hadamard coding, especially for iris

encoding.

During registration, some redundant information (also called helper

data) about the biometric will also be collected and stored to correct the

inconsistent bits.

In almost every published paper on biometric key generation and bind-

ing, the main algorithm(s) can be ascribed to one or a combination of the

above-listed four methods.

BIOMETRIC KEY GENERATION AND BINDING

Biometrics has its origins in the Greek language and is composed of the two

words “bio” (life) and “metrics” (to measure). Historically, Chinese merchants

in the 14th century stamped children’s palm and footprints on paper with

ink to distinguish one from the other (Chirillo & Blaul, 2003). In the 1880s

Frenchman Alphone Bertillon initiated the study of biometrics by developing

methods of identifying criminals. In 1890s Edward Henry started using ﬁnger-

prints for criminal identiﬁcation and later developed the Henry Fingerprint

Classiﬁcation system (Sodhi & Kaur, 2005). The ﬁrst commercial biometric

system, Identimat, was developed in the 1970s (Miller, 1994), which mea-

sured the shape of the hand and the lengths of the ﬁngers. Advances in

electronics during the last two decades of the 20th century make biometrics

more practical for broad applications. In the United States the events of the

9/11 terrorist attack signiﬁcantly increased the interests in biometrics. Since

then a large number of new biometric technologies have been developed.

The main objectives of biometric recognition are user convenience and

better security. I believe that wider applications of biometric technologies are

inevitable and necessary as our society demands higher security. However,

biometric applications have raised a series of issues. Among them the security

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 111

and privacy issue of biometric information is considered more prominent

than other issues. And the solution to the issue is cryptography (i.e., biometric

information needs to be protected with cryptography).

However, cryptography itself has one important issue to be solved: key

management (Goth, 2007). According to “Kerckhoffs’ Principle: The security

of a cryptosystem must not depend on keeping secret the crypto-algorithm.

The security depends only on keeping secret the key” (Reprinted from Singh,

2000). “Key management is the hardest part of cryptography and often the

Achilles’ heel of an otherwise secure system” (reprinted from Schneier, 1996,

p. xxi).

Some researchers have considered the possibility of using biometrics

to solve the key management issue of cryptography. A few methods have

been proposed to correlate a cryptographic key with biometrics. First, Bodo

(1994) proposed to use biometric raw image directly as a cryptographic key.

However, two problems make it non-practical: variations in biometric images

due to uncontrollable environmental and physiological factors make them

not consistent enough to be used as keys and the compromise of a crypto-

graphic key will compromise the privacy of the biometric. Second, biometrics

will be used as a locking mechanism to restrict the access of a cryptographic

key generated with PRNG (Pseudo Random Number Generator)–Biometric

Key Binding (BKB). In BKB the generation of the cryptographic key is in-

dependent of biometrics. The cryptographic key is combined with biometric

information algorithmically at enrollment. Upon veriﬁcation, only when a

matching biometric is presented can the key be retrieved. Some helper data

might also be stored for error correcting purpose. Third, a cryptographic key

is generated directly from biometrics—Biometric Key Generation (BKG).

This method utilizes algorithms to extract a binary string from a biometric

template, which is more robust (error tolerant) than the template itself and

should have enough randomness to be secure and distinct. Ideally it should

be computationally difﬁcult to recover the original biometric raw image and

template from the binary string and the helper data.

In the literature a few biometrics, including keystroke dynamics, voice,

handwritten signatures, ﬁngerprint, face, iris, DNA, and palmprint, have been

proposed for cryptographic key generation and binding. Although many

proposed BKG/BKB algorithms could be applied to different biometrics, I

chose to categorize these proposals based on the types of biometrics for

the following two reasons: (1) generally, one biometric has signiﬁcantly dif-

ferent characteristics from another biometric. Therefore different processing

techniques are needed to transform the raw images into templates; (2) the

majority of published papers report their testing results based only on one

biometric. Under each biometric category I strove to select one representa-

tive paper to provide some detailed information. For other papers I tended

to concisely describe new methods or the new developments. The interested

reader should consult the original paper for more details.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

112 Q. Gao

Keystroke Dynamics

Among other information, a computer user’s typing patterns consist of dura-

tions for each letter typed and latencies between keystrokes, which can be

useful for identifying the user.

Gaines et al. (1980) reported their results on using interkey latencies

to recognize the typists. Obaidat and Sadoun (1997) applied several neural

network and pattern recognition algorithms for verifying computer users

as they type their password phrases and reported that durations are more

effective than latencies in recognizing users.

Monrose et al. (1999) proposed hardening a user’s password with

keystroke dynamics. When a user types his or her password an m-bit string

is derived from the typing and then combined with the password to form

a so-called hardened password. The system is designed in such a way that

while the algorithm runs, the distinguishing features for each account de-

velop over time. Sharp changes in a user’s typing pattern will result in login

failure. The reported results are based on a password which has eight char-

acters, so m=15. Login measurements were recorded for six months of 20

users with at least ﬁve successful logins on their usual keyboards. Similar to

the scheme proposed by Obaidat and Sadoun (1997), the results show that

keystroke durations are more variable among users than keystroke latencies

and that the system has very low key entropy (12 bits) and very high False

Rejection Rate (FRR) (48%).

Bartlow and Cukic (2006) extended the work of Monrose et al. (1999)

by additionally incorporating the shift-key patterns on the changing case of

a letter. With more participants and a larger database, their reported results

show lower False Acceptance Rate (FAR) (3∼6%) and FRR (5∼6%).

Hocquet et al. (2007) proposed a two-step authentication process based

on keystroke. For two successive key strokes, four different times are ex-

tracted:

•PP (Press—Press)

•PR (Press—Release)

•RP (Release—Press)

•RR (Release—Release)

For a sequence of strokes, the system extracts a feature vector for each type

of time. Based on the learning set, each user is associated with a vector of

four features (PP, PR, RP, and RR). K-means a clustering algorithm is used

to cluster all the users into different clusters. Both the typing pattern and the

cluster parameters of a user will be used for recognition.

In sum, keystroke dynamics can be used for text-based password en-

hancement. It has yet to be seen as an effective biometric for cryptographic

key generation or binding.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 113

Voice

Voice is one of the publically accepted behavior biometrics for identiﬁcation.

A few researchers have proposed to generate cryptographic keys from

speech.

Monrose et al. (2001a, 2001b, 2002) proposed a voice-based biometric

system. A speaker’s utterance of an 11-digit telephone number is captured

and divided into a sequence of 30 ms windows with 10 ms overlapping.

A frame with 12 coefﬁcients will be derived from each window. The frame

sequence of the speaker will be segmented into subsequences that are then

matched against a speaker-independent model with 20 different centroids.

The segmentation is a two-step iterative process converging to an optimal

segmentation of the user’s voiceprint: (1) for each segment, ﬁnd the centroid

that yields the maximum likelihood score; (2) use a Viterbi algorithm to

calculate a new segmentation that maximize the product of the segment

likelihood scores relative to the centroid chosen in step 1. Upon completion

of the segmentation process one feature descriptor bit will be generated from

each subsequence. The relative position of the point obtained by averaging

the frames in a subsequence to the closest matching centroid determines

whether the bit is 1 or 0. The reported results are based on 50 people of

both sex and varying nationalities. Each gives ﬁve varying quality utterances

of 1-800-882-3606 over a phone. Two utterances were used for training and

a 46-bit key was extracted from each. The other three were used to test the

false negative rate, which is about 20%.

Garcia-Perera et al. (2005a, 2005b) proposed a procedure for crypto-

graphic key generation from voice, based on the phoneme features of dif-

ferent users using the Support Vector Machines technique for classiﬁcation.

Garcia-Perera et al. (2005c) further improved their scheme by additionally

applying the mixture of Gaussians of the Hidden Markov Model to reﬁne the

classiﬁcation.

Delivasilis and Katsikas (2006) carried out side-channel analysis on

voice-based key generation algorithms in resource constrained devices. With

Differential Power Analysis, they can extract the biometric information used

for key generation and the implementation of a key generation algorithm on

the device.

In sum I believe voice can be a viable biometric for cryptographic

key generation and binding. However, more research needs to be done to

achieve this viability. Researchers with experience in speech recognition may

be able to contribute to a new breakthrough.

Handwritten Signatures

Handwritten signatures as a behavioral biometric are being used for identi-

ﬁcation and key generation.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

114 Q. Gao

Hao and Chan (2002) made use of handwritten signatures to generate

symmetric keys. They deﬁned 43 signature features, such as pen-down time,

max forward Vx (Velocity in x direction), max backward Vy (velocity in y

direction), time when the last peak of Vx or Vy occurs, pressure, height-to-

width ratio, and so on. Ten signatures will be collected from each user to

calculate the mean and standard deviation of the user for each of the 43

features. With 25 users they create a database for each feature. The entire

value range of each feature in the database will be cut into a number of

segments with equal length and each segment will be assigned a different

decimal number D. The value of a particular feature of a particular user will

be one of the assigned decimal numbers. The binary representation for this

feature is obtained by taking the base 2 logarithm of the decimal number (i.e.,

log2D). The bit representations of all the features for a user are concatenated

to form the key. The reported experiments achieved on average a 40-bit key

entropy (a feature may have to be discarded if it is out of boundary) with a

28% false rejection rate and a 1.2% false acceptance rate.

Hoque et al. (2008) proposed a scheme for key generation by partition-

ing feature space into subspaces that are further divided into cells, where

each cell subspace contributes to the overall key generation.

Ballard et al. (2006) illustrated with a generative model that a skilled

human can carry out an effective attack against a handwriting-based key-

generation system with forgeries. To overcome the vulnerabilities Ballard

et al. (2008a) proposed an approach using “pseudo-signatures,” which are

composed of simple shapes like circles, squares, and triangles and are similar

to graphical passwords (Jermyn et al., 1999; van Oorschot & Thorpe, 2007).

Freire-Santosa et al. (2006) studied online signature-based key genera-

tion schemes with different levels of quantization. The experimental results

based on the fuzzy vault (Juels & Sudan, 2002) show that both skilled forg-

eries and random forgeries can have a lower FAR and FRR.

Yip et al. (2006; 2005) proposed schemes on how to generate a re-

placeable and refreshable cryptographic key from dynamic handwritten

signatures.

Kholmatov and Yanikoglu (2006) proposed a scheme that extracts minu-

tiae points (trajectory crossings, endings and points of high curvature) from

online signatures and uses these points to build the fuzzy vault to protect

the 128-bit Advanced Encryption Standard (AES).

In sum I believe cryptographic key generation and binding with hand-

written signature may be subject to forgery because of the inadequacy of

entropy in handwritten signatures.

Palmprint

Human palmprint contains uniqueness and relatively stable features such

as principal lines, wrinkles, minutiae, delta points, area/size of palm, which

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 115

can be used for identiﬁcation purposes. Even identical twins have different

palmprints. Scientists also know that palm lines are associated with some

genetic diseases (Kong et al., 2009).

Palmprints have been proposed for cryptographic key binding. Kumar

and Kumar (2008) proposed an automated key binding system based on

the fuzzy vault of palmprint. Wu et al. (2007) proposed a palmprint key

binding scheme by using a Gaussian Derivative for feature extraction and RS

code for error correction (Refer to Table 2 for more details). Based on the

results from the two reports it seems that palmprints are a good candidate

for cryptographic key generation and binding.

Face

The human face is one of the most widely used biometrics.

“The most popular approaches to face recognition are based on either

the location and shape of facial attributes, such as the eyes, eyebrows, nose,

lips, and chin and their spatial relationships, or the overall (global) analysis

of the face image that represents a face as a weighted combination of a

number of canonical faces” (reprinted from Jain et al., 2004c, p. 9).

A few researchers proposed using face images to generate a crypto-

graphic key.

Goh and Ngo (2003) and Teoh et al. (2004b) outlined cryptographic key-

computation from two-dimensional frontal speaking face bit-map images,

which were taken at a ﬁxed camera distance with 180 ×200 pixels, 8-

bit grayscale, and uniform illumination with dark background. Each image

was then represented as a vector. Principle Component Analysis (PCA) was

then utilized to reduce the dimensions (less than 100 eigenbasis). From

153 individuals, 3,060 images were collected. Ten of 20 images from each

individual were used to construct the eigenbasis for the population. In their

experiments, 20 to 80 bits were extracted from each image. The results

showed that 40 to 60 eigenfaces gives the best results with 0% FA and <3%

FR. Twenty bits is insufﬁcient and 80 bits is sensitive to noise. Chang et al.

(2004) did similar studies with different image processing techniques: PCA,

cascaded Linear Discriminant Analysis (cLDA), and Generalized Symmetric

Max Minimal Distance in Subspace (GSMMS) and concluded that LDA and

GSMMS can improve FAR and FRR due to the more distinguishable feature

transformations.

Chen and Chandan (2007) proposed a key generation method using

iterative, chaotic, bi-spectral one-way transformation for 3D facial feature

extraction. Reed-Solomon error correcting code and a lookup table will be

used for key regeneration.

Zeng and Watters (2007) proposed an approach to generate a key from

a face image by fusing PCA-transformed global feature vectors with Gabor

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

116 Q. Gao

wavelet-transformed local feature vectors. Feature discretization and helper

data are applied for error correction and key renewability.

Zhang et al. (2004) analyzed the security of key generation from face

images with three attacks: Exhaustive Search, Authentic Key Statistics, and

Device Key Statistics, and concluded that a user-dependent system is more

secure than a user-independent one because the former uses different feature

vectors and different thresholds for all the users.

From these research results I believe that the face is a good candidate

for key generation and binding. However, face image is generally regarded

as public information. Therefore, a private key based feature extraction and

image transformation are essential for secure key generation and binding.

DNA

DNA (deoxyribonucleic acid) contains the genetic information of an indi-

vidual. DNA has a double helix structure. Each helix is a 3-dimensional

arrangement of four types of nucleotides or bases: A (adenine), C (cytosine),

G (guanine), and T (thymine), linked by a phosphous-sugar backbone. As

the two helixes are complementary, A always pairs with T and G with C

through Hydrogen bonds. The sequences of the bases, overall more than 3

billion long, determine all the genetic attributes of a person.

The ﬁrst widely accepted DNA analysis technique is the Restriction Frag-

ment Length Polymorphism (RFLP) analysis. It needs a large amount of a

clean DNA sample and takes days if not weeks to get reliable results. A

currently widely used DNA analysis method is the Short Tandem Repeat

(STR)/Polymerase Chain Reaction (PCR) analysis. It uses a smaller amount

of a sample than RFLP and just takes a few hours or less to complete. The

time-consuming procedure is the PCR process, which is carried out to make

a large number of copies of the selected DNA segments. The goal of DNA

analysis is to determine the numbers of the STRs in about a dozen loca-

tions (called loci) recognized by short DNA segments at various loci that are

unique to each individual—except identical twins.

DNA is one of the most important identiﬁcation tools in many areas,

such as Forensics (Ashcroft et al., 2002) and Biometrics (Schneier, 1999).

Most importantly, DNA could be the only biometric containing no fuzziness.

Itakura and Tsuji (2003) proposed a scheme to embed DNA information

into a private and public cryptographic key. Li et al. (2008a) proposed a

DNA-based key generation scheme. Figureau et al. (2000) and Tanaka et al.

(2005) proposed schemes for cryptographic key distribution based on the

mechanism of DNA primer sequence recognition.

In the United States, the Federal Bureau of Investigation (FBI) uses the

number of STRs in 13 positions from 23 pairs of human chromosomes to

identify people. Large amounts of other information can also be extracted

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 117

from DNA. All of the information can be transformed into cryptographic

keys. I expect to see more research to be carried out on how to use people’s

genetic information to protect their digital information.

Iris

Iris is the distinctly colored ring surrounding the pupil of the eye. Made

from elastic connective tissue, the iris is a very rich source of biometric

data, having approximately 266 distinctive characteristics. These include

the trabecular meshwork, a tissue that gives the appearance of dividing

the iris radially, with striations, rings, furrows, a corona, and freckles. Iris

recognition technology uses about 173 of these distinctive characteristics.

Formed during the 8th month of gestation, these characteristics reportedly

remain stable throughout a person’s lifetime, except in cases of injury. Iris

recognition systems use a small, high-quality camera to capture a black

and white, high-resolution image of the iris. The systems then deﬁne the

boundaries of the iris, establish a coordinate system over the iris, and

deﬁne the zones for analysis within the coordinate system. (reprinted

from Vacca, 2007, p. 29)

Although measuring an iris is more intrusive than measuring some other

biometrics like a ﬁngerprint, the iris is regarded as one of the most stable

and accurate biometrics. Many proposals have been made to generate a key

from or bind a key with an iris template.

Hao et al. (2005) proposed a practical and secure way to bind an iris

biometric with a cryptographic key. A key generated with PRNG is XORed

and therefore locked with genuine iris codes at enrollment. To extract the

key, the live iris has to be measured again to get the iris code and then

XORed again with the result obtained from the previous step. To deal with

the 10 to 20% of error bits within an iris code and derive an error-free key,

they carefully studied the error patterns within iris codes, and devised a two-

layer error correction technique that combines Hadamard and Reed-Solomon

codes (MacWilliams & Sloane, 1981). The iris database used consists of 700

iris samples from 70 different eyes, with 10 samples per eye. By varying the

Hadamard codeword lengths (the codeword lengths of Reed-Solomon code

will change correspondingly), keys with different lengths can be locked.

One representative result is to use a 6-bit Hadamard codeword to generate

a 140-bit key with 0.47% FRR and 0% FAR.

Kanade et al. (2009) proposed a two-factor scheme to generate high

entropy cryptographic keys with an iris code and a password. A hash value

of the reference iris code is the key. As in (MacWilliams & Sloane, 1981)

Reed-Solomon codes and Hadamard Codes are used to correct the non-

reproducible bits from iris measurements. Based on a key an iris code is

shufﬂed block wisely to increase its distinctiveness. Also, the iris code is

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

118 Q. Gao

zero-inserted to increase the capacity of Hadamard error correction. Pass-

word controlled AES is applied to protect the ﬁnal enrollment template.

Upon providing the genuine iris sample and the password, the reference

iris code and the cryptographic key can be regenerated. Similarly, Gong et

al. (2008) proposed combining an iris template and PRNG to generate cryp-

tographic keys. Kanade et al. (2008) also proposed a three-factor scheme

to generate biometric cryptographic keys with a smart card, iris code, and

password. They regenerated a 198-bit key with an entropy of 83 bits on the

NIST-ICE database at 0.055% False Acceptance Rate (FAR) and 1.04% False

Rejection Rate (FRR).

Lee et al. (2007, 2008) proposed an iris-based fuzzy vault system to bind

cryptographic key. Pattern clustering and Independent Component Analysis

are used to extract iris code. A shift-matching technique was applied to match

iris templates.

Wu et al. (2008) proposed to use a 2D Gabor ﬁlter to extract a feature

vector from a quantiﬁed iris image with the assistance of Reed-Solomon error

correction. Similar to Kanade (2009) a hash function will be applied to the

vector to generate the cryptographic key.

Castanon et al. (2006) proposed using a generalized regression neural

network (GRNN) to generate a cryptographic key from an iris. The iris is

located with integro-differential operators and features are extracted with

Gabor ﬁltering as introduced by Daugman (2002). The features are then

divided into groups, each of which will provide one input to the GRNN for

key generation.

A binary string generated from biometrics can be used directly as

the key for private-key cryptography, but it may not be so for public

key cryptography, such as RSA (Rivest, Shamir, & Adleman), since the

public–private key pair has to meet certain conditions. Therefore, it is nec-

essary to transform the raw binary string obtained from an iris for public-key

cryptography.

Janbandhu and Siyal (2001a, 2001b) proposed a method of using a 512

byte iris template to generate a private key by ﬁnding the closest larger

number, which is relatively prime with the Euler Totient Function, and using

it as the decryption exponent d of RSA (refer to Table 1). If the number

representing the iris template itself happens to be relative prime to Euler

Totient Function, then use the template directly as the private key. Otherwise,

increment it to get the key.

Mohammadi and Abedi (2008) proposed an ECC-based (Elliptic Curve

Cryptography) biometric signature scheme for PKI (Public-key Infrastructure)

key management, which has some advantages over the RSA-based biometric

signature and makes it possible to generate public and private keys without

storing and transmitting any private information.

In sum the distinctive and rich information content of the iris image

makes it the best candidate for cryptographic key generation and binding.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 119

TABLE 1 Key Generation Based on Iris Code for RSA Cryptosystem

Cryptographic Key Generation/Binding with Iris Code

Operations Explanation/Calculation

Select p, q P and q are prime and 256 byte long, p = q

Calculate n =p×q n is the modulus

Calculate (n) =(p-1) ×(q-1) (n) is the Euler Totient Function

Generate decryption key dfrom the

512-byte iris template by

incrementing it to the closest

number relatively prime to (n)

GCD((n), d) =1

d is the private key

Calculate e e =d−1(mod (n)), e is the public key

However, it is necessary to use a key-based feature extraction to generate

the iris template for a high security requirement.

Fingerprint

Fingerprints have been one of the most widely used and accepted biomet-

rics. Many proposals of biometric key generation and binding are based on

ﬁngerprints.

Soutar et al. (1998a, 1998b; Nichols, 1999) link and retrieve a digital key

by using a ﬁngerprint. At enrollment,4∼6 ﬁngerprint images of a ﬁnger are

collected, averaged, and then multiplied with a random phase-only array. The

phase-phase product of the result, termed Stored Filter Function is saved as

the ﬁrst part of the BioscryptTM. The inverse of the result, a combined image,

will be used to link with an N-bit random number. The linking algorithm

will generate a Look-up Table, which is the second part of the BioscryptTM.

The ﬁrst S bits of the Filter Function will be encrypted with the key; the

encryption result will then be hashed to create an identiﬁcation code, which

is the third part of the BioscryptTM.Atveriﬁcation, a few ﬁngerprint images of

the same ﬁnger are captured, Fourier-Transformed, and averaged to generate

the magnitude information, which is combined with the saved Filter Function

and then inversely Fourier-Transformed to generate a new image, which will

be used with the stored Look-up Table to retrieve the key. Majority decoding

(Davida et al., 1999) and image shifting are the main procedures proposed

for error tolerance.

Song et al. (2008) proposed an application-speciﬁc key release scheme

by using Fingerhashing (Teoh et al., 2004a), Reed-Solomon error correction,

and on-card processing.

Lalithamani and Soman (2009a) proposed a scheme to generate cryp-

tographic key from cancelable ﬁngerprint minutiae (ridge ending and bifur-

cation) template, which was divided into two equal parts and then shufﬂed

to enhance irrevocability. Lalithamani and Soman (2009b) proposed another

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

120 Q. Gao

scheme of using large random numbers to irrevocably deform ﬁngerprint

minutiae, which are then employed for the extraction of keys.

Han et al. (2006) proposed a system to generate cryptographic keys

from ﬁngerprints over public channel. A private-key cryptography system is

used to transmit the biometric image while a public-key cryptosystem is used

to protect the private key, which can be generated directly from the query

biometric images by pixel permutation and gray-level value substitution. The

encryption key for the pixel position shufﬂing is generated by combining the

random pixel distribution of a ﬁngerprint image and features of minutiae-

based ﬁngerprint indexing. The gray-level value modiﬁcation is based on an

extended chaotic baker map.

Han et al. (2007) proposed a reliable way of PIN (Personal Identiﬁcation

Number) generation from ﬁngerprints. First obtain a ﬁngerprint minutiae

template, draw a triangle whose sides connect two minutiae points closest

to the core of a ﬁngerprint image. The longest side, the smallest angle and

the medial angle, together with the minutiae types of the triangle, are used

for the PIN generation.

Song and Teoh (2007) proposed a key extraction scheme. The ﬁnger-

print image is ﬁrst partitioned into a few 16 by 16 blocks, and multi-resolution

log-Gabor ﬁlters with different orientation are applied to obtain the magni-

tude features. Dynamic Quantization Mechanism is used to transform the

magnitude information into a bit string.

Juels and Sudan (2002) proposed a fuzzy vault scheme. The procedure

for constructing the fuzzy vault is as follows: First, Alice selects a polynomial

pof variable xthat encodes k(e.g., by ﬁxing the coefﬁcients of paccording

to k). She computes the polynomial projections, p(A), for the elements of set

A, then adds some randomly generated chaff points that do not lie on p,to

arrive at the ﬁnal point set R. When Bob tries to learn k using his own set

B, he will be able to locate many points in Rthat lie on pif Boverlaps with

Asubstantially to reconstruct p. Otherwise, he will end up with an incorrect

p, and will not be able to access the secret k. The security of this scheme is

based on the infeasibility of the polynomial reconstruction problem.

Clancey et al. (2003) proposed a ﬁngerprint vault based on the fuzzy

vault of Juels and Sudan (2002). They used the canonical position (x, y)

of ﬁngerprints minutia from ﬁve images of a pre-aligned ﬁngerprint and a

polynomial to calculate a set Aand then add a number of chaff points to

get a larger set R. They conclude that a 69-bit key could be obtained with

20∼30% FR and 0% FA. Uludag & Jain (2004) and Uludag et al. (2005) further

improved the fuzzy vault scheme by introducing Cyclic Redundancy Check

(CRC) bits in the secret-containing polynomial. Each ﬁngerprint minutiae in

set A is translated to lie in a square tessellation of a 2D image.

Based on the single biometric controlled fuzzy vault (Juels & Su-

dan, 2002), Hirschbichler et al. (2008) and Hirschbichler and Boyd (2008)

proposed a multiple biometrics controlled fuzzy vault for better security.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 121

Minutiae from multiple ﬁngerprints will be combined in one vault. One or

more polynomials can be selectively applied to the minutiae from different

ﬁngerprints in the vault. Since the ﬁngerprints in the vault can be obtained

from different people, the multiple biometrics controlled fuzzy vault can be

used as scheme for secret sharing. Unlike the original fuzzy vault proposed

by Juels and Sudan (2002), the minutiae angle is also concatenated with the

coordinates.

Al-Tarawneh et al. (2009) proposed a key binding scheme using a

polynomial-based ﬁngerprint fuzzy vault. The entropy of the key can be

controlled by varying the levels of shielding and the degree of the polyno-

mial.

Dodis et al. (2004) proposed an error-tolerance fuzzy extractor that

extracts uniform randomness from biometrics and allows an exact recon-

struction of the original input given an approximate new input with the help

of some extracted public information. They constructed the fuzzy extractor

with hamming distance, set difference, and edit difference. Their work pro-

vides a theoretical guidance for securely generating cryptographic key from

biometrics.

Khalil et al. (2008) proposed a scheme to extract a key from the lo-

cal features of a ﬁngerprint using fuzzy extractor deﬁned in Dodis et al.

(2004).

Boyen (2004) analyzed the insecurity of the schemes proposed in Juels

and Sudan (2002), Juels and Wattenberg (1999), and Dodis et al. (2004) and

proposed a reusable fuzzy extractor against chosen perturbation attack.

Li et al. (2006) analyzed the fuzzy vault scheme in Juels and Sudan

(2002) and found that the application of the Reed-Solomon error correction

code in the UNLOCK algorithm of the fuzzy vault scheme may not be secure.

Chang et al. (2006) proposed a method to separate 2D ﬁngerprint minu-

tiae points from chaff points in a fuzzy vault based on the observation that the

chaff points tend to have a smaller free area and thus have more neighbor-

ing points. Other fuzzy vault based proposals can be found in the following

references (Chung & Moon, 2007; Feng et al., 2008; Jo et al., 2007; Li et al.,

2008c; Lee & Kwon, 2007; Yang & Verbauwhede, 2005).

Scheirer and Boult (2007) introduced three classes of attacks against

biometric fuzzy vault and biometric encryption: attack via record multiplicity,

surreptitious key-inversion attack, and novel blended substitution attacks.

Van Dijk and Tuyls (2005) proposed a theoretical model consisting of

users, multiple veriﬁcation devices, and a certiﬁcation authority on how to

derive a polynomial number of keys from a single biometric and how to

renew keys in a secure and private manner without additional participation

of the user.

Kamara et al. (2008) heuristically analyzed the security of a secret locking

scheme for biometric key generation based on the Shamir Secret Sharing

notion.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

122 Q. Gao

Nguyen and Nguyen (2008) proposed a two-layered cryptographic key

protection scheme: encrypting the to-be-protected key with another key

securely extracted from a ﬁngerprint.

Chen et al. (2007) proposed a scheme for cryptographic key binding

and key retrieval by using a multivariate linear equation.

Izumi et al. (2007) and Izumi and Ueshige (2007) proposed to use a

smart card for biometrics key generation. Helper data for biometric matching

is stored on the card and matching is carried out on the card. No biometric

template is stored to retrieve the key protected by Trusted Authority.

In sum we believe that the ﬁngerprint is going to be one of the ﬁrst few

biometrics used for practical key generation and key binding. The ﬁngerprint

minutiae based fuzzy vault has attracted a lot of attention for cryptographic

key binding. However, the set of minutiae from a single ﬁngerprint may not

contain enough entropy for strong security.

Multi-Biometrics

Humans have a limited number of directly available biometrics. This limita-

tion can be solved by combining multiple biometrics.

Voderhobli et al. (2006) proposed a scheme of using multi-biometrics

for cryptographic key generation. During enrollment, a user provides all

speciﬁed biometric samples, such as ﬁngerprints, iris, voice, facial image,

and a password. The majority of the subsets of the ﬁve samples would

be used to generate keys. During authentication, typically three or more

of the samples have to be matched to regenerate a key. Or a numerical

value can be assigned to each sample and an overall matching score will

be used to determine which key will be generated. Due to the enrollment

of multi-biometrics the performance of the proposed system could be a

problem. From a security viewpoint, combining multi-biometrics makes it

more difﬁcult for an adversary to guess the biometrics and their feature

spaces.

Costanzo (2007) proposed a scheme of key generation by aggregating

features from multibiometrics along with certain mathematical combinatorial

and permutation.

Recently, multi-biometric recognition systems are gaining attention in

the biometric research community. I expect that more research on multi-

biometric key generation and binding will be seen in the near future.

Comparison of Different Biometrics

In Table 2, I have listed some implementations of BKG and BKB. However,

even for the same biometric it is difﬁcult to compare the results from one

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

TABLE 2 Some Implementations of Biometric Key Generation and Binding

Biometrics Feature Error Correcting FRR % FAR % Database Bit Length Reference

Keystroke Duration, latency Discretization 48 NA∗20 users 12 Monrose et al. (1999)

Voice Cepstrum coefﬁcient Discretization 20 NA 10 users 46 Monrose et al. (2001a)

Signature Dynamic Averaging 28 1.2 25 users 40 Hao and Chan (2002)

Face Eigen-projections Discretization 0∼5.0 0 153 users 20∼80 Goh and Ngo (2003)

Face 3D, Fourier Transform RS code 28∼63 0∼1.2 61 users 112∼208 Chen and Chandran (2007)

Iris Gabor ﬁlter RS code Hadamard 0.76 0.10 60 users 94 Kanade et al. (2009)

Iris Gabor wavelet RS code Hadamard 0∼12 0 70 users 42∼224 Hao et al. (2005)

Fingerprint Minutiae Quantization 0.21 0 100 pairs 128 Uludag et al. (2005)

of images

Fingerprint Minutiae RS code 3 0.24 100 users 128 Nandakumar et al. (2007)

Fingerprint Minutiae RS code 5 0 9 users 47 Nagar and Chaudhury (2006)

Fingerprint Fourier Transform Majority coding NA NA NA NA Soutar et al. (1998a)

Palmprint DCT∗∗ RS code 0 0.4 85 users 300 Kumar and Kumar (2008)

Palmprint Gaussian Derivative RS code 0.7∼0.9 0 386 palms NA Wu et al. (2007)

∗Not Available. ∗∗Discrete Cosine Transformation.

123

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

124 Q. Gao

TABLE 3 Qualitative Evaluation of Different Biometrics for Key Generation and Binding

Biometrics Keystroke Voice Signature Palmprint Face DNA Iris Fingerprint

BKG P F P G G E E G

BKB P G P G G E E G

E: Excellent; G: Good; F: Fair; P: Poor.

report with the results from another because they often use different methods

and different databases. Standardization can help to solve this problem.

As stated in the method section, fuzzy measurement is the central prob-

lem of biometric applications. To effectively correct the fuzzy bits, it is often

required to acquire the biometric images under the same conditions for

both enrollment and veriﬁcation. In fact, for each row in Table 2, only one

database is used: one portion for registration and the remaining portion for

testing.

In Table 3 I give a qualitative comparison of the effectiveness of dif-

ferent biometrics for key generation and binding. Generally, a physiological

biometric is more effective than a behavioral biometric for key generation

and binding.

Today biometric cryptographic key generation and binding are still in

the research stage. More research has to be done to achieve the potential of

biometrics as a security enhancing technology for cryptography.

PROTECTING BIOMETRIC IMAGE/TEMPLATE

Many algorithms have been proposed on how to enhance the security of

biometric key generation and binding. Some of these algorithms have been

mentioned brieﬂy in previous sections.

For the majority of biometric key generation systems proposed in the

literature, a biometric template is extracted from the biometric raw image

and has to be saved in a centralized database or distributed on a smart

card. As Ballard et al. (2008b) summarized, biometric key generators (BKG)

contain an enrollment algorithm and a key-generation algorithm along with

four classes of information: (1) Biometric; (2) Template obtained from the

biometric and for key regeneration; (3) Key generated from the biometric

samples; and (4) Auxiliary information including any public information not

used directly in the key generation but available to an adversary. According

to Ballard et al. (2008b), a BKG is secure if it meets the following three

requirements: (1) Key Randomness: Key generated appears random to an

adversary given the template and auxiliary information; (2) Weak Biometric

Privacy (WBP): An adversary cannot reverse-engineer the biometric given the

template and auxiliary information; (3) Strong Biometric Privacy (SBP): An

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 125

adversary cannot reverse-engineer the biometric given the template, auxiliary

information, and the key. Based on the three requirements, they concluded

that the randomness of the key can be signiﬁcantly reduced by the error

correction information and the biometric can be recovered easily by com-

bining the template and auxiliary information. A further compromise of the

key may completely reveal a user’s biometric.

According to Ballard’s deﬁnition of secure BKG, it is extremely difﬁcult,

if not impossible, to securely generate a cryptographic key from biometrics

because of the requirement that a biometric template has to be publicly

available to an adversary. Practically it may be reasonable to relax this re-

quirement. That is to say, a secure BKG can be implemented by preventing

an adversary from accessing the biometric template. Ballard et al. (2008c)

argue that using randomized biometric templates can generate stronger keys

because an adversary has to guess not only the biometric but also how

the biometric is measured. Therefore, the security of a biometric template

becomes a very important issue for secure BKG design.

For the past few years great interest has been paid to protecting biomet-

ric templates (Jain et al., 2008; Nandakumar & Jain, 2008; Sutcu et al., 2007a;

Sutcu et al., 2007b; Sutcu et al., 2007c; Zhou, 2007; Nagar et al., 2008; Feng

et al., 2008b; Maiorana et al., 2008; Xu et al., 2008; Boult et al., 2007; Sun et al.,

2007; Li et al., 2008b; Schipani & Rosenthal, 2007; Vatsa et al., 2006; Linnartz

& Tulys, 2003; Tuyls et al., 2004a; Goseling & Tuyls, 2004; Ang et al., 2005;

Teoh & Ngo, 2005; Pang et al., 2004; Han & Teoh, 2005; Pang et al., 2005;

Teoh & Connie, 2006; Teoh et al., 2004c; Connie et al., 2003a; Connie et al.,

2003b; Lumini & Nanni, 2007; Lumini & Nanni, 2005; Maio & Nanni, 2005;

Davida et al., 1998; van der Veen et al., 2006; Kevenaar et al., 2005; Ratha

et al., 2001; Ratha et al., 2006; Ratha et al., 2007; Tuyls & Goseling, 2004;

Tuyls et al., 2004b; Linnartz & Tuyls, 2003; Verbitskiy et al., 2003; Hidano

et al., 2008; Khan et al., 2009; Draper et al., 2007a; Draper et al., 2007b). Typ-

ically there are two goals in securing a biometric template: First is to achieve

the one-way transformation from image to template. Given a template, one

should not be able to obtain the original biometric image in order to protect

the privacy of the biometric. Second is to generate multiple independent

biometric templates from one image in order to reuse the biometric.

Ratha et al. (2001) formally deﬁned the concept of cancelable biomet-

rics and proposed a few techniques (Image morphing, block permutation,

and polynomial projection) of transforming a biometric template to enhance

the security and privacy of biometrics. Ratha et al. (2006, 2007) also deﬁned

the requirements for template transformation and proposed using Carte-

sian, Polar, and surface folding transformations for ﬁngerprint minutiae-based

templates. To match with transformed templates directly the transformation

needs to tolerate intra-user variability; that is to say, the transformation should

not have a strong diffusion property, deﬁned by Shannon (1949). Due to the

weakness, Feng et al. (2008a) found that it is entirely possible to recover

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

126 Q. Gao

the original minutiae template. The interested reader may refer to two good

reviews (Jain et al., 2008; Cavoukian & Stoianov, 2007) for more details.

Most recently there are new developments in this area that deserve

special attention. Lee et al. (2007) proposed a one-time template by changing

the templates at the completion of every authentication. Gao et al. (2008)

proposed a fuzzy vault based scheme to secure a biometric template by

adding chaff points to biometric templates. However, unlike the fuzzy vault

schemes proposed in Juels and Sudan (2002) in which chaff points have

to be somehow recognized and separated from the real minutiae set later

on, they carried out matching without removing the chaff points. Teoh and

Kim (2007) proposed a biometric Fuzzyhashing scheme striving to achieve

good confusion and diffusion properties as those of a cryptographic one-

way hash function. Sheng et al. (2008) propose a novel method that can

directly generate keys from statistical features of biometric images by using

a fuzzy genetic clustering algorithm and by selecting the most consistent

feature subsets and/or single features for each user individually. Bringer

and Chabanne (2008) and Bringer et al. (2009) propose a new identiﬁcation

method using encrypted biometrics. Nandakumar and Jain (2008) proposed

a scheme to derive a single multi-biometric template from a few individual

templates by transforming features from different biometrics into a common

representation and then performing a feature-level fusion. Fuzzy vault is used

to secure the ﬁnal templates. The experiments show better results than a

uni-biometric vault. Tulyakov (2007) proposed using a high-order symmetric

hash function to transform biometric templates based on minutiae triplets in

order to defend against a Hill-climbing attack (Hill, 2002).

In sum, the security of BKG depends on the privacy of the biometric

image and/or its template. Therefore, protecting a biometric image and/or

template is an inseparable part of secure biometric key generation and

binding.

BIOMETRIC RANDOM NUMBER GENERATOR (BRNG)

Random number generators play a crucial role in information security

(Rukhin et al., 2008). They can be classiﬁed into two types: Truly Ran-

dom Number Generator (TRNG) and Pseudo-Random Number Genera-

tors (PRNG). Most cryptographic keys are generated with software-based

random number generators, which are PRNG. Since computers are de-

terministic ﬁnite-state machines, they cannot produce truly random out-

puts. TRNG utilizes nondeterministic physical sources, such as the posi-

tion/velocity of an atomic electron, decay of a radioactive material, and

quantum entanglement, to produce random numbers. Some researchers pro-

posed using biometrics as an entropy source for random number generation.

Szczepanski et al. (2004) proposed an interesting scheme of using biometric

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 127

information for generating random bit sequences. The two biometrics tested

are Neurophysiological Brain Response and Galvanic Skin Response. Chan-

dran and Chen (2006) proposed a scheme for simultaneous biometric veriﬁ-

cation and generation of a random number that can be used as a privacy key.

Peyravian et al. (1999) proposed algorithms that can generate cryptographic

keys and random numbers from a user’s unique biometric or other user-

speciﬁc information.

I would like to point out that BKG and BRNG are two different instru-

ments. BKG uses biometrics as a type of fuzzy-but-reproducible human iden-

tiﬁcation tool, while BRNG uses biometrics as a random (non-reproducible)

information source.

SUMMARY AND FUTURE RESEARCH

On one hand, biometrics needs to be protected with cryptography. On the

other hand biometrics can help to improve the security of cryptography in

key management. Even though advances in biometrics for authentication and

identiﬁcation have been signiﬁcant, applying biometrics in cryptography is in

its research stage. It is a challenging problem to securely and effectively gen-

erate a cryptographic key from biometrics due to the fuzzy bits in biometric

measurements.

A few error correcting methods, including Averaging/Training, Dis-

cretization, Majority voting, and Error correction coding, have been applied

to approach the non-exact reproducible problem. Even with these methods

it is still required to acquire the biometric images under the same conditions

for both enrollment and veriﬁcation to process the images with the same

algorithms.

A few biometrics including keystroke dynamics, voice, handwritten sig-

nature, palmprint, DNA, ﬁngerprint, and iris, have been proposed for crypto-

graphic key generation and binding. Fingerprint minutiae based fuzzy vault

has attracted a lot of attention for cryptographic key binding. However, the

set of minutiae from a single ﬁngerprint may not contain enough entropy

for strong security. Presently the rich information content makes the iris the

best biometric for BKG/BKB. To date BKG/BKB are still in its early research

stage. Due to the lack of standardization it is difﬁcult to effectively compare

the results from different authors. Generally, a physiological biometric is

more effective than a behavioral biometric for key generation and binding.

The security of BKG depends on the privacy of the biometric image and

its template. Protecting the biometric image and/or template is an inseparable

part of secure biometric key generation and binding.

The security of cryptography relies on efﬁcient random number genera-

tion. Biometrics can also be used for random number generation. However,

BKG and BRNG are two different instruments. BKG uses biometrics as a

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

128 Q. Gao

type of fuzzy-but-reproducible human identiﬁcation tool, while BRNG uses

biometrics as a random (non-reproducible) information source.

In the future, more biometrics, such as DNA, hand geometry, and so

on will be tested for key generation and binding. A multi-biometrics key

generation system will be designed. Another possible research direction is to

develop new algorithms to extract more reliable cryptographic keys as very

high-resolution imaging sensors become available.

REFERENCES

Al-Tarawneh, M., Woo, W., & Dlay, S. (2009). Biometric Key Capsulation Tech-

nique Based on Fingerprint Vault: Anatomy and Attack. 3rd IEEE International

Conference on Information and Communication Technologies: From Theory to

Application, pp. 1–5.

Ang, R., Safavi-Naini, R., & McAven, L. (2005). Cancelable Key-Based Fingerprint

Templates. Lecture Notes in Computer Science (LNCS),3574, 242–252.

Armstrong, I. (2003). Passwords exposed: Users are the weakest link. SC Magazine.

Ashcroft, J., Daniels, D., & Hart, S. (2002). Use DNA to solve cold case. Available at:

http://www.ncjrs.gov/pdfﬁles1/nij/194197.pdf

Ballard, L., Chen, J., Lopresti, D., & Monrose, F. (2008a). Biometric Key Generation

Using Pseudo-Signatures. Proc. of the 11th International Conference on Frontiers

in Handwriting Recognition.

Ballard, L., Kamara, S., & Reiter, M. (2008b). The Practical Subtleties of Biomet-

ric Key Generation. Proc. of the 17th Annual USENIX Security Symposium,

pp. 61–74.

Ballard, L., Kamara, S., Monrose, F., & Reiter, M. (2008c). Towards Practical Biomet-

ric Key Generation with Randomized Biometric Templates. Proc. of 15th ACM

Conference on Computer and Communications Security, pp. 235–244.

Ballard, L., Monrose, F., & Lopresti, D. (2006). Biometric Authentication Revisited:

Understanding the Impact of Wolves in Sheep’s Clothing. Proc. of the 15th

Annual USENIX Security Symposium, pp. 29–41.

Bartlow, N., & Cukic, B. (2006). Evaluating the Reliability of Credential Hardening

Through Keystroke Dynamics. Proc. of the 17th International Symposium on

Software Reliability Engineering. Raleigh, North Carolina, USA.

Bodo, A. (1994). Method for Producing a Digital Signature with Aid of a Biometric

Feature. German patent DE 42 43 90 A1.

Boult, T., Scheirer, W., & Woodworth, R. (2007). Revocable Fingerprint Biotokens:

Accuracy and Security Analysis. IEEE Conference on Computer Vision and Pat-

tern Recognition, pp. 1–8.

Boyen, X. (2004). Reusable Cryptographic Fuzzy Extractors. Proc. of the 11th ACM

Conference on Computer and Communications Security, pp. 82–91.

Bringer, J., & Chabanne, H. (2008). An authentication protocol with encrypted bio-

metric data. LNCS,5023, 109–124.

Bringer, J., Chabanne, H., & Kindarji, B. (2009). Identiﬁcation with Encrypted Bio-

metric Data Made Feasible. International Conference on Communication and

Information Systems Security Symposium.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 129

Castanon, L., Reigosa, M., & Nolazco-ﬂores, J. (2006). Biometric-iris random key

generator using generalized Regression neural networks. LNCS,4031, 530–539.

Cavoukian, A., & Stoianov, A. (2007). Biometric encryption: A positive-sum technol-

ogy that achieves strong authentication, security and privacy. Information and

Privacy Commissioner of Ontario.

Chandran, V., & Chen, B. (2006). Simultaneous Biometric Veriﬁcation and Random

Number Generation. 5th Workshop on Internet, Telecommunications and Signal

Processing.

Chang, E., Shen, R., & Teo, F. (2006). Finding the Original Point Set Hidden among

Chaff. Proc. of the ACM Symposium on Information, Computer and Communi-

cations Security, pp. 182–188.

Chang, Y., Zhang, W., & Chen, T. (2004). Biometric-Based Cryptographic Key Gen-

eration. IEEE Conference on Multimedia and Expo,Taiwan.

Chen, B., & Chandran, V. (2007). Biometric Based Cryptographic Key Generation

from Faces. Proc. of the 9th Biennial Conference of the Australian Pattern

Recognition Society on Digital Image Computing Techniques and Application,

pp. 394–401.

Chen, H., Sun, H., & Lam, K. (2007). Key Management Using Biometrics. 1st

International Symposium on Data, Privacy, and E-Commerce, pp. 321–

326.

Chirillo, J., & Blaul, S. (2003). Implementing Biometric Security, 1st edition. Wiley.

Chung, Y. & Moon, K. (2007). Biometric Certiﬁcate Based Biometric Digital Key

Generation with Protection Mechanism. Proc. of the Frontiers in the Convergence

of Bioscience and Information Technologies, pp. 709–714.

Clancy, T., Kiyavash, N., & Lin, D. (2003). Secure Smartcard-Based Fingerprint Au-

thentication. Proc. of the ACM SIGMM Workshop on Biometrics Methods and

Applications, pp. 45–52.

Connie, T., Teoh, A., Goh, M., & Ngo, D. (2003a). PalmHashing: A novel approach

for dual factor authentication. Pattern Analysis & Applications,7(3), 255–268.

Connie, T., Teoh, A., Goh, M., & Ngo, D. (2003b). PalmHashing: A novel approach

for cancelable biometrics. Information Processing Letter,93(1), 1–5.

Costanzo, C. (2007). Active Biometric Cryptography (ABC): Key Generation Using

Feature and Parametric Aggregation. 2nd International Conference on Internet

Monitoring and Protection.

Daugman, J. (2002). How Iris Recognition Works. Proc. of the International Confer-

ence on Image Processing,1, I33–I36.

Davida, G., Frankel, Y., Matt, B. (1998). On Enabling Secure Applications Through

Off-Line Biometric Identiﬁcation. Proc. of IEEE Symposium on Security and

Privacy, pp. 148–157.

Davida, G., Frankel, Y., Matt, B., & Peralta, R. (1999). On the Relation of Error

Correction and Cryptography to an Off Line Biometric Based Identiﬁcation

Scheme. Proceeding of Workshop on Coding and Cryptography.

Delivasilis, D., & Katsikas, S. (2006). Side channel analysis on biometric-based key

generation algorithms on resource constrained Devices. International Journal

of Network Security,3(1), 44–50.

Dodis, Y., Reyzin, L., & Smith, A. (2004). Fuzzy extractors: How to generate strong

keys from biometrics and other noisy data. LNCS,3027, 523–540.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

130 Q. Gao

Draper, S., Khisti, A., Martinian, E., Vetro, A., & Yedidia, J. (2007a). Using Distributed

Source Coding to Secure Fingerprint Biometrics. MITSUBISHI ELECTRIC RE-

SEARCH LABORATORIES, Technical Report TR2007–005.

Draper, S., Khisti, A., Martinian, E., Vetro, A., & Yedidia, J. (2007b). Secure Storage

of Fingerprint Biometrics Using Slepian-Wolf Codes. MITSUBISHI ELECTRIC

RESEARCH LABORATORIES, Technical Report TR2007-006.

Feldmeier, D., & Karn, P. (1990). UNIX Password Security—Ten Years Later. Proc.

of the Crypto Conference, pp. 44–63.

Feng, Q., Su, F., & Cai, A. (2008). Fingerpring-based key binding/recovering scheme

based on fuzzy vault. Journal of Electronics (China),25(3), 415–421.

Feng, Q., Su, F., Cai, A., & Zhao, F. (2008a). Cracking Cancelable Fingerprint Tem-

plate of Ratha. 2008 International Symposium on Computer Science and Com-

putational Technology,2, 572–575.

Feng, Y., Yuen, P., & Jain, A. K. (2008b). A Hybrid Approach for Face Template

Protection. Proc. of SPIE,6944, 08–11.

Figureau, A., Soto, M., & Toh´

a, J. (2000). Biocryptography. Medical Hypotheses,

54(3), 394–396.

Freire-Santosa, M., Fierrez-Aguilara, J., & Ortega-Garciaa, J. (2006). Cryptographic

Key Generation Using Handwritten Signature. Proc. of SPIE,6202, 225–231.

Gaines, R., Lisowski, W., Press, S., & Shapiro, N. (1980). Authentication by Keystoke

Timing: Some Preliminiary Results”. Rand Report R-256-NSF. The Rand Corpo-

ration, Santa Monica, CA.

Gao, Q., Zhang, X., & Anshel, M. (2008). Experiments on matching intronized ﬁn-

gerprint minutiae templates. International Journal of Computer Science and

Network Security,8(9), 206–213.

Garcia-Perera, L., Mex-Perera, J., & Nolazco-Flores, J. (2005a). Multi-Speaker Voice

Cryptographic Key Generation. ACS/IEEE International Conference on Computer

Systems and Applications (AICCSA’05), pp. 93–I.

Garcia-Perera, L., Mex-Perera, J., & Nolazco-Flores, J. (2005b). Cryptographic-speech-

key generation using the SVM Technique over the lp-Cepstral Speech Space.

LNCS,3445, 370–375.

Garcia-Perera, L., Nolazco-Flores, J., & Mex-Perera, J. (2005c). Cryptographic-speech-

key generation architecture improvements. LNCS,3523, 579–585.

Goh, A., & Ngo, D. (2003). Computation of cryptographic keys from face biometrics.

LNCS,2828, 1–13.

Gong, Y., Deng, K., & Shi, P. (2008) PKI Key Generation Based on Iris Features.

International Conference on Computer Science and Software Engineering,6,

166–169.

Goseling, J., & Tuyls, P. (2004) Information Theoretic Approach to Privacy Protection

of Biometric Templates. Proc. of the International Symposium on Information

Theory.

Goth, G. (2007). Key Management Standards Hit the Fast Track. IEEE Distributed

Systems Online,8(9).

Han, F., Hu, J., & Yu, X. (2006). A biometric encryption approach incorporating

ﬁngerprint indexing in key generation. LNCS,4115, 342–251.

Han, F., Hu, J., He, L., & Wang, Y. (2007). Generation of reliable PINs from

ﬁngerprints. IEEE International Conference on Communication, pp. 1191–

1196.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 131

Han, N., & Teoh, A. (2005). A novel cancelable face authentication biometrics. Spe-

cial Issue on Pattern Recognition in Biometrics and Bioinformatics, Multimedia

CyberSpace Journal.

Hao, F., & Chan, C. (2002). Private key generation from on-line handwritten signa-

tures. Information Management & Computer Security,10(2), 159–164.

Hao, F., Anderson, R., & Daugman, J. (2005). Combining cryptography with biomet-

rics effectively. Technical Reports, University of Cambridge, Computer Labora-

tory.

Hidano, S., Ohki, T., Komatsu, N., & Kasahara, M. (2008). On Biometric Encryption

Using Fingerprint and it’s Security Evaluation. 10th International Conference on

Control, Automation, Robotics and Vision, pp. 950–956.

Hill, C. (2002). Risk of Masquerade Arising from the Storage of Biometrics. Thesis

for the Bachelor of Science, Dept. of CS, Australian National University, 2002.

Hirschbichler, M., & Boyd, C. (2008). A Multiple Controlled Fuzzy Vault. 6th Annual

Conference on Privacy, Security and Trust, pp. 36–47.

Hirschbichler, M., Boles, W., & Boyd, C. (2008). A Scheme for Enhancing Security

Using Multiple Fingerprints and the Fuzzy Vault. Proc. of Digital Image Com-

puting: Techniques and Applications, pp. 540–547.

Hoque, S., Fairhurst, M., & Howells, G. (2008). Evaluating Biometric Encryption

Key Generation Using Handwritten Signatures. Proc. of the 2008 Bio-inspired,

Learning and Intelligent Systems for Security, pp. 17–22.

Hocquet, S., Ramel, J., & Cardot, H. (2007) User classiﬁcation for keystroke dynamics

authentication. LNCS,4642, 531–539.

Itakura, Y., & Tsuji, S. (2003). Proposal on personal authentication system in

which biological information is embedded in cryptosystem key. Information

Processing Society of Japan,Information Processing Society of Japan,45, 19–

27.

Izumi, A., Ueshige, Y., & Sakurai, K. (2007). A proposal of key management scheme

and its operation using anonymous biometrics on ID-based infrastructure. In-

ternational Journal of Security and Its Applications,1(1), 83–94.

Izumi, A., & Ueshige, Y. (2007). A Proposal of Efﬁcient Scheme of Key Management

Using ID-Based Encryption and Biometrics. 2007 International Conference on

Multimedia and Ubiquitous Engineering, pp. 29–34.

Jain, A. K., Dass, S., & Nandakumar, K. (2004a). Soft biometric traits for personal

recognition systems. LNCS,3072, 731–738.

Jain, A. K., Hong, L., & Pankanti, S. (2000). Biometrics: Promising Frontiers for

Emerging Identiﬁcation Market. Communications of the ACM, pp. 91–98.

Jain, A. K., Hong, L., Pankanti, S., & Bolle, R. (1997). An Identity Authentication

System Using Fingerprints. Proc. of IEEE,85(9), 1365–1388.

Jain, A. K., Nandakumar, K., Lu, X., & Park, U. (2004b). Integrating faces, ﬁn-

gerprints, and soft biometric traits for user recognition. LNCS,3087, 259–

269.

Jain, A. K., Nandakumar, K., & Nagar, A. (2008). Biometric template security. Journal

on Advances in Signal Processing, Special Issue on Advanced Signal Processing

and Pattern Recognition Methods for Biometrics.

Jain, A. K., Ross, A., & Prabhakar, S. (2004c). An Introduction to Biometric Recog-

nition. IEEE Transactions on Circuits and Systems for Video Technology, Special

Issue on Image- and Video-Based Biometrics,14(1), 4–20.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

132 Q. Gao

Janbandhu, P., & Siyal, M. (2001a). Novel biometric digital signature for Internet

based applications. Information Management and Computer Security,9(5),

205–212.

Janbandhu, P., & Siyal, M. (2001b). A New Biometric Based Signature System. Proc.

of the 3rd International Conference on Information, Communication and Signal

Processing.

Jermyn, I., Mayer, A., Monrose, F., Reiter, M., & Rubin, A. (1999). The Design and

Analysis of Graphical Passwords. Proc. of the 8th USENIX Security Symposium.

Jo, J., Seo, J., & Lee, H. (2007). Biometric digital signature key generation and

cryptography communication based on ﬁngerprint. LNCS,4613, 38–49.

Juels, A., & Sudan, M. (2002). A Fuzzy Vault Scheme. IEEE International Symposium

on Information Theory.

Juels, A., & Wattenberg, M. (1999). A Fuzzy Commitment Scheme. Proc. of ACM

Conference on Computer and Communications Security, pp. 28–36.

Kamara, S., Medeiros, B., & Wetzel, S. (2008). Secret locking: Exploring new ap-

proach to biometric key encapsulation. Communications in Computer and In-

formation Science,3, 101–112.

Kanade, S., Camara, D., Krichen, E., Petrovska-Delacretaz, D., & Dorizzi, B. (2008).

Three Factor Scheme for Biometric-Based Cryptographic Key Regeneration Us-

ing Iris. 6th Biometrics Symposium.

Kanade, S., Camara, D., Petrovska-Delacretaz, D., & Dorizzi, B. (2009). Application

of Biometrics to Obtain High Entropy Cryptographic Keys. Proc. of the World

Academy of Science, Engineering and Technology, 39.

Kevenaar, T., Schrijen, G., Van Der Veen, M., Akkermans, A., & Zuo, F.

(2005). Face Recognition with Renewable and Privacy Preserving Binary Tem-

plates., 4th IEEE Workshop on Automatic Identiﬁcation Advanced Technologies,

pp. 21–25.

Khalil, M., Muhammad, D., & Ahmed, M. (2008). Converting Fingerprint Local Fea-

tures to Public Key Using Fuzzy Extractor. The International Arab Conference

on Information Technology.

Khan, M., Xie, L., & Zhang, J. (2009) Chaos and NDFT-based spread spectrum

concealing of ﬁngerprint-biometric data into audio signals. Digital Signal

Processing.

Kholmatov, A., & Yanikoglu, B. (2006). Biometric cryptosystem using online signa-

tures. LNCS,4263, 981–990.

Klein, D. V. (1990) Foiling the Cracker: A Survey Of, and Improvements To, Password

Security. Proc. of the 2nd USENIX Workshop Security, pp. 5–14.

Kong, A., Zhang, D., & Kamel, M. (2009). A survey of palmprint recognition. Pattern

Recognition,42, 1408–1418.

Kumar, A., & Kumar, A. (2008). A Palmprint-Based Cryptosystem Using Double

Encryption. Proc. of SPIE,6944,1–9.

Lalithamani, N., & Soman, K. (2009a). An effective scheme for generating irrevocable

cryptographic key from cancelable ﬁngerprint templates. International Journal

of Computer Science and Network Security,9(3), 183–193.

Lalithamani, N., & Soman, K. (2009b). Irrevocable cryptographic key generation from

cancelable ﬁngerprint templates: An enhanced and effective scheme. European

Journal of Scientiﬁc Research,31(3), 372–387.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 133

Lee, Y., Bae, K., Lee, S., Park, K., & Kim, J. (2007). Biometric key binding: Fuzzy

vault based on iris images. LNCS,4642, 800–808.

Lee, Y., Bae, K., Lee, S., Park, K., & Kim, J. (2008). A new method for generat-

ing an invariant iris private key based on the fuzzy vault system. IEEE Trans-

actions on Systems, Man, and Cybernetics. Part B, Cybernetics,38(5), 1302–

1313.

Lee, Y., Chung, Y., & Moon, K. (2007). One-Time Templates for Face Authenti-

cation. International Conference on Convergence Information Technology, pp.

1818–1823.

Lee, H., & Kwon, T. (2007). Biometric digital key mechanisms for telebiometric

authentication based on biometric certiﬁcate. LNCS,4554, 428–437.

Li, Q., Liu, Z., & Niu, X. (2006). Analysis and Problems on Fuzzy Vault Scheme.

International Conference on Intelligent Information Hiding and Multimedia

Signal Processing, pp. 244–250.

Li, X., Zhang, L., & Hu, Y. (2008a). A Novel Generation Key Scheme Based on

DNA. International Conference on Computational Intelligence and Security, 1,

264–266.

Li, Y., Yin, J., Zhu, E., Hu, C., & Chen, H. (2008b). Score Based Biometric Tem-

plate Selection and Update. 2nd International Conference on Future Generation

Communication and Networking,3, 35–40.

Li, J., Yang, X., Tian, J., Shi, P., & Li, P. (2008c). Topological Structure-Based Align-

ment for Fingerprint Fuzzy Vault. 19th International Conference on Pattern

Recognition, pp. 1–4.

Linnartz, J., & Tulys, P. (2003) New shielding functions to enhance privacy and

prevent misuse of biometric templates. LNCS,2688, 393–402.

Linnartz, P., & Tuyls, P. (2003). New Shielding Functions to Enhance Privacy and

Prevent Misuse of Biometric Templates. International Conference on Audio-

and Video-Based Biometric Person Authentiﬁcation,2688, 393–402.

Lumini, A., & Nanni, L. (2005). Empirical tests on biohashing, NeuroComputing,

69(16), 2390–2395.

Lumini, A., & Nanni, L. (2007). An improved biohashing for human authenticaiton.

Pattern Recognition,40(3), 1057–1065.

MacWilliams, F., & Sloane, N. (1981) The Theory of Error-Correcting Codes. North

Holland.

Maio, D., & Nanni, L. (2005). MultiHashing, human authentication featuring biomet-

rics data and tokenised random number: a case study FVC2004. NeuroComput-

ing,69, 242–249.

Maiorana, E., Martinez-Diaz, M., Campisi, P., Ortega-Garcia, J., & Neri, A. (2008).

Template Protection for HMM-based On-line Signature Authentication. IEEE

Computer Society Conference on Computer Vision and Pattern Recognition

Workshops, pp. 1–6.

Maltoni, D., Maio, D., Jain, A. K., & Prabhakar, S. (2003) Handbook of Fingerprint

Recognition (1st Edition).Springer.

Miller, B. (1994). Vital signs of identity. IEEE Spectrum,31(2), 22–30.

Mohammadi, S., & Abedi, S. (2008). ECC-Based Biometric Signature: A New Ap-

proach in Electronic Banking Security. International Symposium on Electronic

Commerce and Security, pp. 763–766.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

134 Q. Gao

Monrose, F., Reiter, M., & Wetzel, S. (1999). Password Hardening Based on Keystroke

Dynamics. Proc. of the ACM Conference in Computer and Communications

Security, pp. 73–82.

Monrose, F., Reiter, M., Li, Q., & Wetzel, S. (2001a). Cryptographic Key Generation

From Voice. Proc. of the IEEE Symposium on Security and Privacy.

Monrose, F., Reiter, M., Li, Q., & Wetzel, S. (2001b). Using voice to generate crypto-

graphic keys. Proc. of Odyssey 2001, The Speaker Veriﬁcation Workshop.

Monrose, F., Reiter, M., Li, Q., Lopresti, D., & Shih, C. (2002). Toward Speech

Generated Cryptographic Keys on Resource Constrained Devices. Proc. of the

11th USENIX Security Symposium, pp. 283–296.

Nagar, A., Nandakumar, K., & Jain, A. K. (2008). Securing Fingerprint Template:

Fuzzy Vault with Minutiae Descriptors. Proc. of International Conference on

Pattern Recognition, pp. 1–4.

Nagar, A., & Chaudhury, S. (2006). Biometrics Based Asymmetric Cryptosystem De-

sign Using Modiﬁed Fuzzy Vault Scheme. Proc. of International Conference on

Pattern Recognition, pp. 537–540.

Nandakumar, K., & Jain, A. K. (2008). Multibiometric Template Security Using Fuzzy

Vault. The 2nd IEEE International Conference on Biometrics: Theory, Applica-

tions and Systems, pp. 1–6.

Nandakumar, K., Jain, A. K., & Pankanti, S. (2007). Fingerprint-Based Fuzzy Vault:

Implementation and Performance. IEEE Transaction on Information Forensics

and Security, pp. 744–757.

Nichols, R. (1999). ICSA (International Computer Security Association) Guide to Cryp-

tography. Chapter 22. McGraw-Hill.

Nguyen, T., & Nguyen, T. (2008). An Approach to Protect Private Key Us-

ing Fingerprint Biometric Encryption Key in BioPKI Based Security System.

10th International Conference on Control, Automation, Robotics and Vision,

pp. 1595–1599.

O’Gorman, L. (2003). Comparing Passwords, Tokens, and Biometrics for User Au-

thentication. Proc. of IEEE,91(12), 2019–2040.

Obaidat, M., & Sadoun, B. (1997). Veriﬁcation of Computer Users Using Keystroke

Dynamics. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cyber-

netics,27(2), 261–269.

Pang, Y., Teoh, A., & Ngo, D. (2004). Cancelable palmprint authentication system.

International Journal of Signal Processing,1(2), 98–104.

Pang, Y., Teoh, A., & Ngo, D. (2005). Binarized revocable biometrics in face recog-

nition. LNCS,3802, 788–795.

Peyravian, M., Matyas, S., Roginsky, A., & Zunic, N. (1999). Generating user-based

cryptographic keys and random numbers. Computers & Security,18(7), 619–626.

Ratha, N., Chikkerur, S., Connell, J., & Bolle, R. (2007). Generating Cancelable Fin-

gerprint Templates. IEEE Transactions on Pattern Analysis and Machine Intelli-

gence,29(4), 561–572.

Ratha, N., Connell, J., & Bolle, R. (2001). Enhancing security and privacy in

biometrics-based authentication systems. IBM Systems Journal,40(2), 614–634.

Ratha, N., Connell, J., Bolle, R., & Chikkerur, S. (2006). Cancelable Biometrics: A Case

Study in Fingerprints. 18th International Conference on Pattern Recognition,

pp. 370–373.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 135

Rukhin, A. et al. (2008). A Statistical Test Suite for the Validation of Random Number

Generators and Pseudo Random Number Generators for Cryptographic Appli-

cations. NIST Special Publication 800-22 (Rev.1).

Scheirer, W., & Boult, T. (2007). Cracking Fuzzy Vault and Biometric Encryption.

Proc. of Biometrics Symposium.

Schipani, D., & Rosenthal, J. (2007). Coding solutions for the secure bio-

metric storage problem. Available at: http://arxiv.org/PS cache/cs/pdf/

0701/0701102v1.pdf

Schneier, B. (1996). Applied Cryptography (2nd Edition). John-Wiley, New York.

Schneier, B. (1999). Biometrics: Uses and Abuses. Communications of ACM,42(8).

Shannon, C. (1949). Communication theory of secrecy Systems. Bell System Technical

Journal,28(4), 656–715.

Sheng, W., Howells, G., Fairhurst, M., & Deravi, F. (2008). Template-Free Biometric

Key Generation by Means of Fuzzy Genetic Clustering. IEEE Transaction on

Information Forensics and Security,3(2), 183–191.

Singh, S. (2000). The Code Book: The Science of Secrecy from Ancient Egypt to Quan-

tum Cryptography, p. 12. Anchor Books, New York.

Skaff, G. (2007). An alternative to passwords? Biometric Technology Today,15(5),

10–11.

Sodhi, J., & Kaur, J. (2005). The forgotten Indian pioneers of ﬁngerprint science.

Current Science,88(1), 185–191.

Song, O., & Teoh, A. (2007). Fuzzy Key Extraction from Fingerprint Biometrics

Based on Dynamic Quantization Mechanism. 3rd International Symposium on

Information Assurance and Security, pp. 71–76.

Song, O., Teoh, A., & Chek, D. (2008). Application-Speciﬁc Key Release Scheme

from Biometrics. International Journal of Network Security,6(2), 127–

133.

Soutar, C., Roberge, D., Stoianov, A., Gilroy, R., & Kumar, B. (1998a). Biometric

EncryptionTM—Enrollment and Veriﬁcation Procedures. Proc. of SPIE,3386,

24–35.

Soutar, C., Roberge, D., Stoianov, A., Gilroy, R., & Kumar, B. (1998b). Biometric

EncryptionTM Using Image Processing. Proc. of SPIE,3314, 178–188.

Sun, S., Lu, C., & Chang, P. (2007). Biometric Template Protection: A Key-

Mixed Template Approach. International Conference on Consumer Electronics,

pp. 1–2.

Sutcu, Y., Li, Q., & Memon, N. (2007a). Protecting Biometric Templates with Sketch:

Theory and Practice. IEEE Transactions on Information Forensics and Security,

2(3), 503–512.

Sutcu, Y., Li, Q., & Memon, N. (2007b). Secure Biometric Templates from Fingerprint-

Face Features. Proc. of the IEEE Computer Society Conference on Computer Vision

and Pattern Recognition, pp. 1–6.

Sutcu, Y., Sencar, H., & Memon, N. (2007c). A Geometric Transformation to Protect

Minutiae-Based Fingerprint Templates. Proc. of SPIE, 6539.

Szczepanski, J., Wajnryb, E., Amig´

o, J., Sanchez-Vives, M., & Slater, M. (2004). Bio-

metric random number generators. Computers & Security,23(1), 77–84.

Tanaka, K., Okamoto, A., & Saito, I. (2005) Public-key system using DNA as a

one-way function for key distribution. Biosystems,81(1), 25–29.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

136 Q. Gao

Teoh, A., & Connie, T. (2006). Remarks on biohashing based cancelable biometrics

in veriﬁcation system. NeuroComputing,69(16-18), 2461–2464.

Teoh, A., & Kim, J. (2007). FuzzyHash: A Secure Biometric Template Protection

Technique. Frontiers in the Convergence of Bioscience and Information Tech-

nologies, pp. 688–694.

Teoh, A., & Ngo, D. (2005). Cancelable Biometrics Featuring With Tokenised Ran-

dom Number. Pattern Recognition Letter,26(10), 1454–1460.

Teoh, A., Ngo, D., & Goh, A. (2004a). FingerHashing: Two factor authentication

featuring ﬁngerprint data and tokenized random number. Pattern Recognition,

37, 2245–2255.

Teoh, A., Ngo, D., & Goh, A. (2004b). Personalised cryptographic key generation

based on Facehashing. Computers & Security,23, 606–614.

Teoh, A., Ngo, D., & Goh, A. (2004c). Biohashing: Two factor authentication featuring

ﬁngerprint data and tokenised random number. Pattern Recognition,37(11),

2245–2255.

Tulyakov, S., Farooq, F., Mansukhani, P., & Govindaraju, V. (2007). Symmetric hash

functions for secure ﬁngerprint biometric systems.Pattern Recognition Letters,

28(16), 2427–2436.

Tuyls, P., & Goseling, J. (2004). Capacity and examples of template protecting bio-

metric authentication systems. LNCS,3087, 158–170.

Tuyls, P., Skoric, B., & Kevenaar, T. (2007). Security with Noisy Data: On Private

Biometrics, Secure Key Storage and Anti-Counterfeiting, 1st Edition. Springer.

Tuyls, P., Verbitskiy, E., Goseling, J., & Denteneer, D. (2004a). Privacy protecting

biometric authentication systems: An overview. Philip Research. Available at:

http://www.eurasip.org/content/Eusipco/2004/defevent/papers/cr1921.pdf

Tuyls, P., Verbitskiy, E., Ignatenko, T., Schobben, D., & Akkermans, A. (2004b).

Privacy Protected Biometric Templates: Acoustic Ear Identiﬁcation. Proc. of SPIE,

5404, 176–182.

Uludag, U., & Jain, A. K. (2004). Fuzzy Fingerprint Vault. Proc. of Workshop on

Biometrics: Challenges Arising from Theory to Practice, pp. 13–16.

Uludag, U., Pankanti, S., & Jain, A. K. (2005). Fuzzy Vault for Fingerprints. Proc. of

Audio and Video-based Biometric Person Authentication, pp. 310–319.

Vacca, J. (2007). Biometric Technologies and Veriﬁcation Systems, 1st edition.

Butterworth-Heinemann.

van Der Veen, M., Kevenaar, T., Schrijen, G., Akkermans, A., & Zuo, F. (2006). Face

Biometrics with Renewable Templates. Proc. of SPIE, 6072.

van Dijk, M., & Tuyls, P. (2005). Secure Biometrics. Available as of May 2, 2009 at:

http://www.eurasip.org/Proc./Eusipco/Eusipco2005/defevent/papers/cr1193.

pdf.

van Oorschot, P., & Thorpe, J. (2007). On Predictive Models and User-Drawn Graph-

ical Passwords. ACM Transactions on Information and System Security.

Vatsa, M., Singh, R., Noore, A., Houck, M., & Morris, K. (2006). Robust biometric

image watermarking for ﬁngerprint and face template protection. The Institute

of Electronics, Information and Communication Engineers, Electronics Express,

3(2), 23–28.

Verbitskiy, E., Tuyls, P., Denteneer, D., & Linnartz, J. (2003). Reliable Biometric Au-

thentication with Privacy Protection. 24th Benelux Symposium on Information

Theory.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Recent Developments on Applying Biometrics in Cryptography 137

Voderhobli, K., Pattinson, C., & Donelan, H. (2006). A Schema for Cryptographic

Keys Generation Using Hybrid Biometrics. 7th Annual Postgraduate Sympo-

sium: The Convergence of Telecommunications, Networking and Broadcasting.

Liverpool, UK.

Wu, X., Qi, N., Wang K., & Zhang, D. (2008). A Novel Cryptosystem Based on Iris Key

Generation. Proc. of the 4th International Conference on Natural Computation,

04, 53–56.

Wu, X., Zhang, D., & Wang, K. (2007). A palmprint cryptosystem. LNCS,4642,

1035–1042.

Xu, W., He, Q., Li, Y., & Li, T. (2008). Cancelable Voiceprint Templates Based on

Knowledge Signatures. International Symposium on Electronic Commerce and

Security, pp. 412–415.

Yang, S., & Verbauwhede, I. (2005). Automatic Secure Fingerprint Veriﬁcation Sys-

tem Based on Fuzzy Vault Scheme. Proc. of IEEE International Conference on

Acoustics, Speech, and Signal Processing,5, 609–612.

Yip, W., Goh, A., Ngo, D., & Teoh, A. (2005). Refreshable cryptographic key gener-

ation from dynamic handwritten signature. Special issue on Pattern Recognition

in Biometrics and Bioinformatics, Multimedia CyberScape Journal.

Yip, W., Goh, A., Ngo, D., & Teoh, A. (2006). Generation of replaceable crypto-

graphic keys from dynamic handwritten signatures. LNCS,3832, 509–515.

Zeng, Z., & Waters, P. (2007). A Novel Face Hashing Method with Feature Fusion

for Biometric Cryptosystems. Proc. of the 4th European Conference on Universal

Multiservice Networks, pp. 439–444.

Zhang, W., Zhang, C., & Chen, T. (2004). Security Analysis for Key Generation

Systems Using Face Images. IEEE International Conference on Image Processing,

pp. 3455–3458.

Zhou, X. (2007). Template Protection and its Implementation in 3D Face Recognition

Systems. Biometric Technology for Human Identiﬁcation IV, Proc. of SPIE,6539,

214–225.

Downloaded By: [Gao, Qinghai] At: 03:10 10 September 2010

Applications of Digital Signature Certificates for Online Information Security

Chapter

May 2018

M. Tariq Banday

Information security has been the focus of research since decades; however, with the advent of Internet and its vast growth, online information security research has become recurrent. Novel methods, techniques, protocols, and procedures are continuously developed to secure information from growing threats. Digital signature certificates, currently offers one of the most trusted solutions to achieve CIAtrio for online information. This chapter discusses online information security through cryptography. It explains digital signature certificates; their benefits, the underlying standards, involved techniques, procedures, algorithms, processes, structure, management, formats, and illustration of their working. It highlights the potential of digital signatures and certificates in information security across different devices, services, and applications. It introduces a few useful tools to learn, train, and implement digital signature certificates.

Applications of Digital Signature Certificates for Online Information Security

Chapter

Oct 2015

M. Tariq Banday

Information security has been the focus of research since decades; however, with the advent of Internet and its vast growth, online information security research has become recurrent. Novel methods, tech- niques, protocols, and procedures are continuously developed to secure information from growing threats. Digital signature certi cates, currently o ers one of the most trusted solutions to achieve CIA- trio for online information. This chapter discusses online information security through cryptography. It explains digital signature certi cates; their bene ts, the underlying standards, involved techniques, procedures, algorithms, processes, structure, management, formats, and illustration of their working. It highlights the potential of digital signatures and certi cates in information security across di erent devices, services, and applications. It introduces a few useful tools to learn, train, and implement digital signature certi cates.

Privacy Friendly Biometrics

Conference Paper

Full-text available

Apr 2015

A wider use of biometric devices has led to a greater level of privacy loss. Numerous organizations have created centralized databases of individuals, which often go beyond the legal framework. When such data are compromised, it can mean a kind of " digital death " of an individual. On the other hand, there is an increasing need for true random generated sequences (TRNG) that can serve as a unique crypto key of individuals. This paper incorporates the concept known аs " revocable " or " cancellable biometrics " and offers schemes (in this particular case of traffic control) where biometric data become the specific guardian of one's personal identity, under a term " Privacy friendly biometrics ".

67-70 - PRIVACY FRIENDLY BIOMETRICS - Milivoj Mrdakovic, Saša Ž Adamović

Data

Full-text available

Jun 2015

A wider use of biometric devices has led to a greater level of privacy loss. Numerous organizations have created centralized databases of individuals, which often go beyond the legal framework. When such data are compromised, it can mean a kind of “digital death” of an individual. On the other hand, there is an increasing need for true random generated sequences (TRNG) that can serve as a unique crypto key of individuals. This paper incorporates the concept known аs “revocable” or “cancellable biometrics” and offers schemes (in this particular case of traffic control) where biometric data become the specific guardian of one’s personal identity, under a term “Privacy friendly biometrics”.

Privacy Friendly Biometrics

Conference Paper

Full-text available

Apr 2015

A wider use of biometric devices has led to a greater level of privacy loss. Numerous organizations have created centralized databases of individuals, which often go beyond the legal framework. When such data are compromised, it can mean a kind of “digital death” of an individual. On the other hand, there is an increasing need for true random generated sequences (TRNG) that can serve as a unique crypto key of individuals. This paper incorporates the concept known аs “revocable” or “cancellable biometrics” and offers schemes (in this particular case of traffic control) where biometric data become the specific guardian of one’s personal identity, under a term “Privacy friendly biometrics”

Application of biometrics to obtain high entropy cryptographic keys

Article

Mar 2009

In this paper, a two factor scheme is proposed to generate cryptographic keys directly from biometric data, which unlike passwords, are strongly bound to the user. Hash value of the reference iris code is used as a cryptographic key and its length depends only on the hash function, being independent of any other parameter. The entropy of such keys is 94 bits, which is much higher than any other comparable system. The most important and distinct feature of this scheme is that it regenerates the reference iris code by providing a genuine iris sample and the correct user password. Since iris codes obtained from two images of the same eye are not exactly the same, error correcting codes (Hadamard code and Reed-Solomon code) are used to deal with the variability. The scheme proposed here can be used to provide keys for a cryptographic system and/or for user authentication. The performance of this system is evaluated on two publicly available databases for iris biometrics namely CBS and ICE databases. The operating point of the system (values of False Acceptance Rate (FAR) and False Rejection Rate (FRR)) can be set by properly selecting the error correction capacity (t s) of the Reed- Solomon codes, e.g., on the ICE database, at t s = 15, FAR is 0.096% and FRR is 0.76%.

Comparing passwords, tokens, and biometrics for user authentication

Article

Dec 2003

J Esch

Using DNA to solve cold cases

Chapter

Jan 2011

Irrevocable cryptographic key generation from cancelable fingerprint templates: An enhanced and effective scheme

Article

Jan 2009

The growing vastness of identity theft in our society has made the requirement of reliable information security mechanisms a priority. Even though information security can be accomplished with the help of a prevailing tool like cryptography, protecting the confidentiality of the cryptographic keys is one of the significant issues to be deal with. This predicament can be efficiently solved by the integration of biometrics with cryptography. Of late, the enhanced performance of cryptographic key generated from biometrics in terms of security has obtained massive reputation amongst the researchers and experimenters. However, there is a permanent association between the biometric and the user, in which modification is impossible. Hence, the biometric is gone eternally and possibly for all the applications which apply it, if there is a compromise of the biometric identifier. This may be solved by the construction of revocable biometric templates through cancelable biometrics. In this paper, we have proposed an approach to generate irrevocable cryptographic key from cancelable fingerprint templates, which performs in an efficient manner. Initially, the fingerprints are employed to extract the minutiae points which are transformed in an efficient manner to obtain deformed points. Subsequently, the deformed points are employed to generate the cancelable templates which are utilized for the extraction of irrevocable keys. It is highly impractical to acquire cancelable fingerprint templates and original fingerprints from the generated key owing to the fact that cryptographic key generated is irrevocable. We have utilized the fingerprints accessible from open sources to evaluate the efficacy of our approach. In addition, the protection analysis of the proposed method has been presented.

Palmprint based cancelable biometric authentication system

Article

Jan 2004

Cryptographic Key Generation Using Handwritten Signature

Article

Jan 2006

Based on recent works showing the feasibility of key generation using biometrics, we study the application of handwritten signature to cryptography. Our signature-based key generation scheme implements the cryptographic construction named fuzzy vault. The use of distinctive signature features suited for the fuzzy vault is discussed and evaluated. Experimental results are reported, including error rates to unlock the secret data by using both random and skilled forgeries from the MCYT database.

Implementing Biometric Security

Book

Jan 2003

Security with noisy data. Private biometrics, secure storage and anti-counterfeiting

Book

Jan 2007

Corporate Performance Management (CPM) is a basic approach which examines the relationship between corporate performance and process optimization. How to successfully introduce CPM in practice is demonstrated through project reports from E.ON, British Telecom, Credit Suisse and Vodafone among others. The methods and tools presented here guarantee a continuous and automated monitoring of the corporate performance and enable Business Process Excellence to be permanently established in the company by company-internal and company-external benchmarking. The articles in this book focus on the use of the ARIS Controlling Platform developed by IDS Scheer.

Score Based Biometric Template Selection and Update

Conference Paper

Dec 2008

A biometric system usually contains two stages: registration and authentication. Most biometric systems capture multiple samples of the same biometric trait at registration. As a result, it is essential to select several samples as templates. This paper proposes two algorithms maximum match scores (MMS) and greedy maximum match scores (GMMS) based on match scores for template selection and update. The proposed algorithms need not involve the specific details about the biometric data. Therefore, they are more flexible and can be used in various biometric systems. The two algorithms are compared with Random and sMDIST on the database of FVC2006DB1A, and the experimental results show that the proposed approaches can improve the accuracy of biometric system efficiently. Based on the maximized score model, we propose two strategies: ONLINE and OFFLINE for templates update. And then we analyze the relationship between them. Preliminary experiments demonstrate that OFFLINE strategy gains better performance.

NIST Special Publication 800-22: A Statistical Test Suite for the Validation of Random Number Generators and Pseudo Random Number Generators for Cryptographic Applications

Article

Apr 2010

This paper discusses some aspects of selecting and testing random and pseudorandom number generators. The outputs of such generators may be used in many cryptographic applications, such as the generation of key material. Generators suitable for use in cryptographic applications may need to meet stronger requirements than for other applications. In particular, their outputs must be unpredictable in the absence of knowledge of the inputs. Some criteria for characterizing and selecting appropriate generators are discussed in this document. The subject of statistical testing and its relation to cryptanalysis is also discussed, and some recommended statistical tests are provided. These tests may be useful as a first step in determining whether or not a generator is suitable for a particular cryptographic application. However, no set of statistical tests can absolutely certify a generator as appropriate for usage in a particular application, i.e., statistical testing cannot serve as a substitute for cryptanalysis. The design and cryptanalysis of generators is outside the scope of this paper. Key words: random number generator, hypothesis test, P-value

Recent Developments on Applying Biometrics in Cryptography

Abstract and Figures

Recommended publications

Ever had your fingerprints taken? Meeting the challenges of 21st Century access control

Biometric template protection: Towards a secure biometric system

Improving the Efficiency of Capture-Resistant Biometric Authentication Based on Set Intersection

Privacy-Aware Biometrics: Design and Implementation of a Multimodal Verification System