ArticlePDF Available

Abstract and Figures

Individuals, businesses and governments undertake an ever-growing range of activities online and via various Internet-enabled digital devices. Unfortunately, these activities, services, information and devices are the targets of cybercrimes. Verifying the user legitimacy to use/access a digital device or service has become of the utmost importance. Authentication is the frontline countermeasure of ensuring only the authorized user is granted access; however, it has historically suffered from a range of issues related to the security and usability of the approaches. They are also still mostly functioning at the point of entry and those performing sort of re-authentication executing it in an intrusive manner. Thus, it is apparent that a more innovative, convenient and secure user authentication solution is vital. This paper reviews the authentication methods along with the current use of authentication technologies, aiming at developing a current state-of-the-art and identifying the open problems to be tackled and available solutions to be adopted. It also investigates whether these authentication technologies have the capability to fill the gap between high security and user satisfaction. This is followed by a literature review of the existing research on continuous and transparent multimodal authentication. It concludes that providing users with adequate protection and convenience requires innovative robust authentication mechanisms to be utilized in a universal level. Ultimately, a potential federated biometric authentication solution is presented; however it needs to be developed and extensively evaluated, thus operating in a transparent, continuous and user-friendly manner.
This content is subject to copyright. Terms and conditions apply.
Cluster Comput
DOI 10.1007/s10586-015-0510-4
Continuous and transparent multimodal authentication:
reviewing the state of the art
Abdulwahid Al Abdulwahid1,2·Nathan Clarke1·Ingo Stengel1·
Steven Furnell1·Christoph Reich3
Received: 20 July 2015 / Revised: 5 November 2015 / Accepted: 7 November 2015
© Springer Science+Business Media New York 2015
Abstract Individuals, businesses and governments under-
take an ever-growing range of activities online and via various
Internet-enabled digital devices. Unfortunately, these activ-
ities, services, information and devices are the targets of
cybercrimes. Verifying the user legitimacy to use/access a
digital device or service has become of the utmost impor-
tance. Authentication is the frontline countermeasure of
ensuring only the authorized user is granted access; however,
it has historically suffered from a range of issues related to
the security and usability of the approaches. They are also
still mostly functioning at the point of entry and those per-
forming sort of re-authentication executing it in an intrusive
manner. Thus, it is apparent that a more innovative, con-
venient and secure user authentication solution is vital. This
paper reviews the authentication methods along with the cur-
rent use of authentication technologies, aiming at developing
a current state-of-the-art and identifying the open problems
to be tackled and available solutions to be adopted. It also
investigates whether these authentication technologies have
the capability to fill the gap between high security and user
satisfaction. This is followed by a literature review of the
existing research on continuous and transparent multimodal
authentication. It concludes that providing users with ade-
quate protection and convenience requires innovative robust
authentication mechanisms to be utilized in a universal level.
BAbdulwahid Al Abdulwahid
1Centre for Security, Communications and Network Research,
Plymouth University, Plymouth PL4 8AA, UK
2Computer Science and Engineering Department, Jubail
University College, Jubail Industrial City,
Kingdom of Saudi Arabia
3Institute for Cloud Computing and IT-Security, Furtwangen
University of Applied Science, Furtwangen, Germany
Ultimately, a potential federated biometric authentication
solution is presented; however it needs to be developed and
extensively evaluated, thus operating in a transparent, con-
tinuous and user-friendly manner.
Keywords User authentication ·Authentication technolo-
gies ·Security ·Usability ·Transparent authentication ·
Biometrics ·Continuous authentication
1 Introduction
Protecting an IT system against unauthorized user activities
is usually provided via user identification or authentica-
tion which enable successful authorization and subsequently
accountability—these concepts together are referred to as
AAA [1]. The identity of a user is required by a system
to authenticate/verify user’s credentials against an authen-
tication database to decide whether he/she is the legitimate
claimed individual. For instance, a username is a way of
claiming an identity and a password is one method for provid-
ing authentication. Proceeding to a successful verification,
authorization is established based on the predefined devices
and/or services the verified user is allowed to access on a
system with specified privileges. Accountability provides the
means to attribute activities each user performs on a system
and keeps tracks of them—usually through historical logs.
Therefore, managing appropriate authentication is the piv-
otal concept for implementing information security within an
IT system. Achieving a high level of confidentiality, integrity,
authorization, and accountability of an IT system would not
be possible without carefully considering various aspects; a
vital one of them is safeguarding sensible, robust and useable
authentication. Authentication can be achieved by utilizing
one or more of the three fundamental approaches: some-
Cluster Comput
thing the user knows (including passwords, PINs, graphical
passwords, and cognitive questions), something the user has
(including SIMs, smart cards, certificates, mobile phones,
and hardware/software one-time password (OTP) tokens)
and something the user is (biometrics) [2].
The first two authentication approaches have been
employed in most security systems surrounding today’s dig-
ital society. However, the third one has emerged gradually
from being research and utilized mainly by governments (e.g.
forensics and borders), to becoming more available in the
public domain (biometrics are now deployed in a wide range
of applications that are fairly mainstream—passports, mobile
phones, schools, police).
The authors aim at building an authentication system that
would provide a more secure, user-friendly, universal, and
technology independent environment. In order to achieve
this, the following research objectives are established:
To review the authentication methods including both the
problems and available solutions.
To investigate the state-of-the-practice of authentication
technologies provided by various sectors.
To develop a current state-of-the-art understanding of the
biometric authentication techniques including its appli-
cations in the existing research on continuous, transparent
and distributed authentication.
This paper is structured as follows: Section 2reviews the
conventional authentication approaches. Then Sect. 3exam-
ines the current use of authentication technologies offered
by service providers and devices manufacturers in order
to explore whether they solve some issues related to the
research area. Furthermore, a number of featured authen-
tication frameworks are subsequently discussed in Sect. 4,
in terms of the benefits they offer in balancing the trade-off
between security and usability as well as their shortcomings.
Furthermore, Sect. 5undertakes a thorough review of the lit-
erature related to continuous and transparent authentication
focusing upon those utilized multimodal biometrics, encom-
passing their open issues, users’ perceptions, and desirable
requirements, leading to an outline of the proposed solution
alongside its limitations and future changes. Finally, the con-
clusion and sought features are presented in Sect. 6.
2 Conventional authentication approaches
2.1 Secret knowledge-based approach
This approach refers to the process where the user has to
remember a secret which is a particular sequence of inputs,
typically made up of numbers only (PIN); numbers, char-
acters and/or symbols (password and passphrase); answer(s)
to predefined question(s) (cognitive knowledge); or images
(graphical password) [3]. This secret is set initially by the user
or generated by the authenticating system. Thus, it is known
mutually by both the user (brain) and the system (database)
and there must be an exact match between them to be able to
have access. This means that it is a Boolean authentication
process—its outcome is either one (totally true secret thus
allow access) or zero (totally false secret thus deny access).
As a result, there is an integral reliance on humans’ mem-
ory and their ability to recall the secret exactly as and when
prompted regardless of its length, sophistication, and unique-
ness. Furthermore, it does not defend well against repudiation
[4] as the so called secret is transferable, guessable and can
be watched by others through shoulder surfing.
2.1.1 Personal identification number (PIN), password and
A PIN is considered the simplest knowledge-based authenti-
cation technique. It is apparently available to be used within
mobile phones: for the mobile handset itself (switch on or
unlock) and/or for the Subscriber Identity Module (SIM)
card (to authenticate with the cellular networks) and with
cash/credit cards. Typically, a mobile PIN ranges from 4 to
8 digits only. As numbers only are relatively easier to recall,
they are easier to guess and to steel. Passwords, which can
be longer and are made of some or all of numbers, letters and
symbols, mitigate the possibility of being predicted. They are
believed to offer effective protection if they are established
and employed appropriately.
Despite the fact that passwords are still the most ubiq-
uitous authentication method (perhaps due to its perceived
convenience and inexpensive implementation as they are
conceptually quite simple to design, manage and use), they
are vulnerable to be misused by users. PINs/Passwords
protections are often compromised through the failure or
unwillingness of individuals to correctly practice the pass-
word policy to protect and administer sensitive information
[5,6]. For instance, 58% of the latter survey respondents
never changed their PINs. Worse than that, it is also revealed
in the former survey of 330 young people aged 18–25 that
over 71 % of the participants do not even use PINs or any other
authentication methods to lock their mobile phones though
their availability. Further more recent survey conducted by
Crawford and Renaud [7] showed that 30% of participants
do not enable any security on their mobile devices although
sensitive information resides on them. Whilst some practice
improvements are notable, the small population (30 partici-
pants) of this survey is an issue but even so when factoring
this percentage to the worldwide mobile users it will be sig-
More recently, many digital services create password poli-
cies and guidelines to encourage good practice, which are
Cluster Comput
adopted by many organizations to be utilized by their employ-
ees. Some of these policies are difficult to ensure they are
being followed and hence they can be avoided. For instance,
it is possible to violate these policies by using dictionary
words, using them on multiple systems, writing them down
and not or rarely changing them. For example 61% of 1200
surveyed respondents reuse the same password on multiple
websites, besides 44 % of them change their password merely
once a year or less [8]. Others are enforceable, such as the
length of password, complexity and its lifetime. Accordingly,
when users are faced with the need to memorize multi-
ple passwords and change them periodically, they tend to
forgetting passwords, writing them down, and selecting eas-
ily guessed ones [4]. Therefore, the problem is exacerbated
as they would become susceptible to be stolen. Moreover,
additional administrative costs would be posed by frequent
passwords resetting [4]. The above-mentioned studies also
implied that some people would rather setup the same but
very sophisticated password on multiple accounts; however
this exasperates the issue if one of these accounts is compro-
mised, all others may follow, as the intruder will be able to
reuse the same cracked password to login to them.
Passphrases come as an alternative endeavor to balance the
trade-off between the simplicity of remembering a secret by
the genuine user and the difficulty of predicting it by intrud-
ers. Passphrases are sequence of words built to be used as
credential secret. They are usually without spaces but pos-
sibly with digits replacing letters or words; for example,
“Going4al0n9journy”. It can be noted that they are similar to
passwords in terms of usage and appearance except that the
former are longer normally thus more robust. On the other
hand, it is argued that passphrases are easier to remember than
passwords especially if they carry an associated meaning.
However, if they consist of common words from a language
dictionary, they would be vulnerable to be broken with less
effort. In addition, common substitutes, such as “4=for” and
“0=o”, render it less secure and more confusing to recall
Brute-force attack tools (attempting every possible com-
bination automatically), such as Brutus and OphCrack, are
notorious against most of knowledge-based authentication
techniques [9]. Some countermeasures have been proposed
against them and to reduce the likelihood of a system or
device being abused by imposters during the usage session
and before it ends. For instance, the account would be tempo-
rary blocked or further credentials would be requested after
three failed access attempts or the user would be required
to re-authenticate again after specific or lapse time depen-
dent upon the system settings or the user’s preference. Even
though that this seems to move the PINs, passwords and
passphrases from being a mere point-of-entry technique, it
most probably bothers the user due to its constant intrusive-
2.1.2 Cognitive knowledge question
Cognitive knowledge which comes in a form of question(s)
seeks to alleviate the load of users memorizing desperate
passwords thereby deploying associative question(s) [10].
These questions are typically about personal information,
such as mother’s maiden name and city of birth, or pref-
erences, such as favorite color and movie. Therefore, it is
evident that this technique lacks one of the main characteris-
tics of secret knowledge-based authentication approach, i.e.
secrecy. By predicting or conducting online search or social
engineering, it is possible to have the correct answer(s)—the
higher the possibility of an answer to deduce or associate,
the higher it is vulnerable to crack.
So, it is apparent that this approach cannot be depend-
able as a standalone authentication approach. This could be
overcome by requiring a user to answer a group of cogni-
tive knowledge questions or alternatively utilizing it besides
another authentication approach (as explained in Sect. 2.4).
Whilst this solution probably enhances security by adding
another layer, it potentially increases the burden on the user
thereby lengthening the time of authentication and requiring
them to recall and provide multiple secrets (i.e. the pass-
word and the answers of the cognitive questions). However,
this approach offers opportunities of supporting the secu-
rity level of other than secret-knowledge ones, such as OTP
tokens. Furthermore, it can be used as a remedial approach
for resetting the password when users for instance forget their
password or are locked-out due to exceeding the maximum
failed login attempts.
2.1.3 Patterns and graphical passwords
Solutions have been suggested to mitigate the downsides of
PIN, password and passphrase, some of which solely concern
about guidelines promoting increasing the entropy of pass-
words. However, human inability to memorize and remember
multi complex passwords is not addressed by them. It is
believed and has been proven that the human brain is more
capable to store and remember pictorial information than tex-
tual [11]. As a result, pattern password authentication has
emerged, with which a user is required to recognize and
sequentially draw a pre-set outline on nine (3 ×3) dots grid
that appear on a touch screen. Therefore, it is argued that
it will be much more convenient to the user to recognize
a pattern than an alphanumeric password. In addition, [12]
showed and argued that repeated entry of pictorial password
would be with “lower cognitive load and higher memora-
bility” to the user. Mobile devices with touch screens make
it reasonably plausible to utilize pattern password, which is
used in Android devices, to improve the memorability of the
Cluster Comput
Fig. 1 Pattern with the possibility of points to be skipped [13]
However, in the current functioning pattern passwords,
users are able only to stroke and drag (draw a direct line
between) two adjacent dots, which in turn limit the number of
permutations. As a result, the typical application of it is more
vulnerable to brute-force attacks. Some attempts have been
conducted to overcome this shortcoming. For instance, [13]
extend this typical pattern password to allow skipping dots (as
demonstrated in Fig. 1), thus enhancing its resilience to brute-
force attacks by allowing more combinations. Nevertheless,
its accuracy is quite low (77 %) with a 19 % false rejection rate
and 21 % false acceptance rate. Furthermore, besides the fact
that this approach is still secret-knowledge based and hence
inherits most of its drawbacks, such as shoulder surfing, it
is susceptible to a so-called smudge attacks when a secret
pattern can be simply determined on a greasy screen [14].
To obtain the most from the advantages of human’s abil-
ity to remember graphical over alphanumeric secrets, some
approaches have been proposed. For example, with click-
based graphical authentication, there is a generic image
where the user is required to click on pre-specified obscured
points [15]. Albeit evaluations have demonstrated its usabil-
ity improvement in relation to memorability, it is relatively
difficult to click precisely on a point, especially if the point
space is small and while using finger tips on touch screen.
This leads to increase authentication failures that might
bother the user. Moreover, poor selection of background
images that have popular potential points yields to being eas-
ily predicted, for instance a study by [16] cracked an average
of 7–10 % of user passpoints (click-based) passwords within
3 guesses only.
Further to the work on click-based concept, proposals
about choice-based or PassImages graphical authentication
have risen [17,18]; in addition to the recent application of the
concept on Windows 8.1 Picture Password [19]. There are a
set of images on sequential grids; the secret is among them in
a form of a series of images that should be pressed or clicked
on a specific order, one at each grid. To overcome shoulder
surfing attacks, the distribution of images on each grid should
be randomized. Likewise, the product of [20] capitalizes on
the psychological theory that human’s brains recognize and
recall faces better than any other picture or object [21]. Users
are able to use familiar personal photos that are stored on the
ones PC or on the web to form passfaces, with which the pos-
sibility of forgetting them is very rare. In the login process,
the user is encountered by a 3 by 3 grid that contains one of
the pre-set photos among 8 others. Similar to the other graph-
ical password methods, there are three consecutive grids to
identifying all three faces. Accordingly, the time taken to pass
all the steps of graphical authentication could be an issue of
inconvenience. Again, poor selection of photos makes them
susceptible to be known by imposters. Moreover, given that
it is a secret-knowledge approach, it can be shared and left
not changed.
2.2 Token-based approach
To overcome some of the abovementioned downsides of
secret knowledge-based approach, tokens have been devel-
oped. Generally, the token-based authentication approach
has various applications ranging from physical to logical
accesses to systems and services. Based on the external
appearance and the need for additional devices, they can be
categorized into two types: Hardware Tokens and Software
Tokens [22]. With the former type, a separate token physi-
cal device is produced and provided, usually, by the service
provider, such as bank smartcard and HSBC Secure Key OTP
token [23]. On the other hand, with the latter type, there is
utilization of an existing device as is, such as when sending
OTP via short messaging service (SMS) to the user’s reg-
istered mobile phone, or there is a need to install software
(application) on the user’s smartphone or PC [22], such as
Google authenticator [24].
A typical authentication token either stores static but com-
plex passwords or generates a OTP for each session [25]. The
user is required to enter the generated password on the system
or service he is authenticating to or it is synchronized directly.
From one prospective, they have some advantages over the
secret knowledge-based methods in that they are capable of
storing, recalling and generating multiple and sophisticated
passwords, thus lifting this burden from the human brain.
However, reliance on human is still existent as it is assumed
that the token is in the possession of the accredited user—they
merely verify the presence of the token not the authorized
user. Having said this, in recent tokens, PIN is prompted to
validate the user for a subsequent legitimate use of the token;
however, the token can be lent, lost or stolen and the PIN can
be shared.
Tokens provide compromise detection, for example if
three failed attempts threshold is exceeded, as well as coun-
termeasure denial-of-service attacks [4], albeit they are not
fail-safe—the breach of RSA SecureID tokens in 2011 evi-
Cluster Comput
dences this [26]. Therefore, it is evident that this approach
cannot stand by itself to be effective at inhibiting masquer-
ade attacks. As a result, typically, it is employed with at
least another authentication factor to form an approach called
multi-factor authentication which is elaborated in the Sect.
It is apparent that the cost of issuing, maintaining and
recovering them is higher. Simply issuing (or reissuing if
lost or stolen) SIM, smart cards or hardware tokens is adding
additional cost over passwords. This is worsened if special-
ized devices are required, such as card readers. For example,
if a bank plans to employ hardware tokens to access its online
banking, there is a need to purchase tokens/token readers
for all its customers, implement and maintain them, along
with providing technical support and potential replacement
in case they are lost or malfunctioned. Moreover, time syn-
chronization between the token and system might be difficult
with those time-synchronous tokens [25], especially in out-
of-coverage areas. Furthermore, user convenience is an issue,
in particular when users need to carry a variety of tokens
for different accounts and services from different providers
which make it cumbersome and probably impractical.
2.3 Biometrics
In seeking a more reliable and robust authentication app-
roach, attention has turned to biometrics. Biometrics-based
authentication is commonly acknowledged as a reliable solu-
tion that provides enhanced authentication over the secret
knowledge-based and token-based approaches. Unlike the
previous approaches, biometrics enables both identification
and verification processes. Regardless of whether the user has
claimed an identity initially or not, the high level of unique-
ness biometrics offers facilitates the process. It also removes
the reliance upon the individual to either memorize and recall
complex and various passwords or carry and secure tokens.
However, whilst the resulting decision of other approaches
is with complete accuracy (i.e. a Boolean decision), biomet-
rics results in a confidence measure, with a pre-determined
threshold deciding on whether this confidence is sufficient to
accept or reject access. Thus, there is a margin for this deci-
sion being wrong; either by allowing access to an imposter
or denying access of the authorized user. Accordingly, the
performance of a typical biometrics technique is measured
based on its error rates, such as False Acceptance Rate (FAR),
False Rejection Rate (FRR) and Equal Error Rate (EER).
Biometrics is dependent upon measurable and distinctive
characteristics of an individual. They can be categorized
based upon their underlying characteristics into: physio-
logical and behavioral approaches [27,28]. Physiological
biometrics are those based upon a unique physical aspect of
the body, such as a fingerprint, face, or iris, whereas behav-
ioral biometrics utilizes the distinctive way in which humans
behave, such as voice, keystroke and signature, to identify
and/or verify a user. Both categories are non-transferable
to others, unforgettable, believed to uniquely (with a vary-
ing level of accuracy) identify individuals, not easily lent or
stolen, and difficult to reproduce, change or hide. As such,
they offer a strong defense against repudiation [29]. How-
ever, biometric systems error rates and cost, together with
usability have been hindering their widespread adoption [30];
notwithstanding, recent years have shown that this has been
alleviated by significant enhancement in biometric systems
capabilities [31,32]. Nevertheless, stable uni-biometrics can
be forged albeit some with difficulty [4]. For instance, tra-
ditional facial recognition can be fooled by a photo of the
authorized person and voice recognition can be faked by
imitation or voice recording. Therefore, they can be used in
combination with a token that can store the user’s identity or a
password (as elucidated in the following sub-section) or addi-
tional data is required to determine whether a sample is alive.
Liveness detection have been suggested and implemented to
determine whether the provided biometric sample is from
a living legitimate user utilizing some biological indicators,
such as blood flow and blinking for iris scan, and temper-
ature and pulse for fingerprint systems [10,25,33]. Whilst
these metrics have added a level of protection, some of them
suffer from their own weaknesses and hence are forgeable.
For instance, an impersonator can hold a photo of an autho-
rized person with two eye holes, stand behind it and blink in
front of a facial recognition system. However, devising a bio-
metrics system deploying a set of countermeasures makes it
robust and difficult to compromise. Alternatively, multibio-
metrics would offer a more resilient authentication solution
as can be seen in Sect. 5.1.
2.4 Multi-layer and -factor authentication
To improve and augment the level of protection, two or more
authentication techniques can be employed in combination. It
has, even, been recommended by the European Central Bank
that financial service providers should deploy “strong authen-
tication” in all their online transactions [34]. It can comprise
multiple techniques from the same authentication approach
(multi-layer authentication), such as password and cogni-
tive questions, or from different authentication approaches
(multi-factor authentication), such as PIN and smart card,
password and facial recognition, or fingerprint and OTP gen-
erator token. This can then be reinforced by elements such
as predefined user location which can be based on either the
mobile cellular network (i.e. cell ID), the global positioning
system (GPS) (i.e. longitude and latitude) [1], and/or the IP
The multi-layer method lack adherence to regulations of
some sensitive sectors, such as banks where it is not com-
patible with the Federal Financial Institutions Examination
Cluster Comput
Council regulations that emphasized clearly that these fac-
tors are required to be from two or more of the authentication
categories [35]. Therefore, it can be inferred that multi-factor
authentication is considered stronger than multi-layer one—
thus banking sector has utilized multi-factor authentication
in one way or another, such as the bank card and PIN or
password and OTP token for online banking. On the other
hand, although some recent smartphones are equipped with
a built-in facial recognition or fingerprint sensor, they oper-
ate separately as an alternative single authentication method
not multi-factor, i.e. the user has the option either to enable
PIN or the fingerprint not both of them together. Hence, to the
author best knowledge no multi-factor authentication method
has been utilized to access mobile phones thus far.
Nevertheless, while the aforementioned approaches
increase the level of security, they add a further burden, from
the perspective of the user, and remain at the point-of-entry.
Re-authenticating the user periodically is not viable because
of its intrusiveness. Furthermore, they increase the cost of
provisioning, managing and implementing various authenti-
cation methods.
3 A review of current use of authentication
It can be perceived that the integral aim of any IT authenti-
cation system is to safeguard resources against any illegiti-
mate access. Therefore, service providers as well as device
manufacturers require or offer a form of authentication tech-
nologies to protect them from any unauthorized access.
Authentication technologies vary perhaps dependent on the
data sensitivity involved and the users’ requirements, and
each have their own benefits and weaknesses. This section
investigates some of the available provided authentication
mechanisms, with the aim of identifying their capabilities
for accomplishing the aim of this research.
A number of service providers and devices manufactur-
ers offer a variety of authentication technologies seeking to
fill the gap between high protection and usability. Thus, it
is useful to review some of these attempts with the current
authentication technologies employed with/by a sample of
service/device providers; namely:
HSBC [23],
NatWest [36],
Lloyds [37],
SAMBA (Saudi American Bank) [38],
Windows 8.1 Laptop/PC [19,39],
Android (Samsung Galaxy S5 and above) [40,41],
iPhone 5S and above [42,43] and
Google Authenticator [24].
This set was selected because it is believed that they represent
a wide range of services and providers that offer a variety
of advanced authentication methods. Moreover, due to the
fact that banks hold high sensitive financial data, they are
expected to strive to deploy the most advanced robust identity
verification procedures. Other less critical and/or less com-
mon service providers and services are deemed not to utilize
such resilient protection tools. Thus, half of the selected list
is banks in addition to the most dominant operating systems
[44]. Google Authenticator is also included for the sake of
diversity and inclusion as it has a different approach than the
remaining listed technologies and it works with many lead-
ing websites such as Amazon Web Services, Dropbox, and
Facebook [45].
Table 1reveals an overview of these authentication tech-
nologies in order to better appreciate whether they have
solved and mitigated the issues of traditional authentication
flaws by enhancing security as well as improving the usabil-
ity of authentication.
Accessing all of the services mentioned in Table 1above
requires a form of secret-based information, including user
ID, PIN, password, pattern, and/or cognitive question(s) all
of which are needed to be memorized and recalled by users.
All of these services except Lloyds bank augment their
authentication process by offering the option of employ-
ing multi-factor authentication or imposing it. To be able
to unlock an Android (Galaxy S5/6) or iPhone (5S) device,
a user selects to provide either a secret (i.e. PIN or pass-
word (for both), pattern (for Android)) or biometrics (i.e.
face/fingerprint, or fingerprint, respectively).
On the other hand, accessing HSBC and SAMBA online
banking systems must happen by entering secret information
(i.e. user ID and cognitive question or password), in addi-
tion to having a separate hardware token for either banks,
or using the user’s mobile as token that generates OTP or
via SMS, respectively. However, two of the services employ
two-layer authentication for the initial access: NatWest and
Lloyds banks. The former asks only for user ID and pass-
word whereas the latter adds them with a cognitive question
to log in. Nevertheless, the user will be prompted to pro-
vide an additional credential, OTP, when a critical service is
requested, such as creating new payee. To do so, NatWest
customers ought to have digital banking card with a sepa-
rate PIN to use with their Card-Reader to generate the OTP
while Lloyds customers will see a OTP on screen and they
will receive an automated phone call to their pre-registered
mobile for confirmation.
These techniques might be perceived as a sensible trade-
off between security and convenience. However, they
arguably on one hand merely augment security but on the
other hand degrade user friendliness, or the vice versa. For
example, with HSBC, NatWest and SAMBA, the user must
carry a separate token which only proves its presence not the
Cluster Comput
Tabl e 1 An overview of current authentication technologies
Secret-based Token-based Biometrics-based Point-of-entry Re-authentication
HSBC [23]User ID Separate
Hardware OTP
X(New OTP)
Cognitive question New payee
PIN Transfer money
NatWest [36]User ID Separate
Hardware OTP
X(New OTP)
PIN Digital
banking card
New payee
Password New standing order
Change password
Change phone
Lloyds [37]User ID (New OTP with
Password X X Automated call to
Cognitive question registered mobile)
New payee
Transfer money
SAMBA [38]User ID Separate
Hardware OTP
X(NewOTP)OR (ATM login)
Password OR New payee
Mobile (SMS)
Transfer money
Windows 8.1
User ID X X Websites accounts
Picture password
Android (Galaxy S5)
PIN X Face X
Pattern Fingerprint
iPhone (5S) [42,43]PIN X Fingerprint Access iTunes
Password New purchase
Google Authenticator
User ID
Mobile OTP X X
legitimacy of the user. Additionally, logging in Lloyds online
banking requires the user to recall 3 distinct secrets. Given
the difficult users experience with remembering secrets and
tokens, these approaches merely serve to increase this bur-
The Google Authenticator app can offer an alternative
solution as it is available in different platforms including
iOS, Android and Blackberry and is easier to use than sepa-
rate tokens as smartphones are carried around by users most
of the time. Conversely, the backup secrets (that can be used
if there is a difficulty in receiving the automatically generated
code) can be stored in the device in an unencrypted text file
[24]. Once it is lost or stolen, the service is susceptible to be
accessed by the unauthorized holder of the device.
On the other hand, there are some encouraging signs and
endeavors regarding classifying the services according to
their level of sensitiveness when prompting re-authentication
to access those ranked higher, such as transferring money
to other accounts, adding a new payee and purchasing from
iTunes. Despite their indication to reflect the reality of fluctu-
ating confidence on the user and services varying risk levels,
should this procedure occur very often, the user is likely to
get bothered.
A few other attempts to utilize biometrics appear with
Windows 8.1, Galaxy S5 and iPhone 5S. For example,
Microsoft declares that they will embed the functionality of
fingerprints to access their apps in Windows 8.1, such as Win-
dows Store, Xbox Music, and Xbox Video [39]. Similarly,
Cluster Comput
Galaxy S5 and iPhone 5S employ the fingerprint scanner on
their home button not only to login but also to access some
apps, such as PayPal and iTunes. Nevertheless, offering the
option of bypassing the fingerprint for PIN or password, even
if they are enabled, may render the feature not being used at
all or render this process to be exploited by attackers where
the drawbacks of secret codes remain.
4 Featured authentication frameworks
A number of researches have upheld the need for more
innovative authentication methods that aim to balance the
trade-off between security and convenience. The following
sub-sections discuss the related two of these featured authen-
tication frameworks, namely single sign-on and federated
identity, in terms of the benefits they offer as well as their
4.1 Single sign-on
An attempt to increase convenience and reduce the burden
(of remembering many passwords and of entering the user’s
credentials on each resource and application) from the user
has evolved–single sign-on (SSO). SSO provides the user
transparent access to all services that they have the privi-
leges to access within an organization after a single successful
login [4,46,47]. They, therefore, only need to set and recall
one password to authenticate to a resource and subsequently
attain the permission to access other services under the same
domain without being prompted to authenticate again. A
popular example is Google account with which the account
holder is required to enter his/her credentials once to be able
to use its services, such as Gmail, Google drive and Google
calendar, during the same session.
Besides the usability benefits from the users perspec-
tives, SSO is perceived to be beneficial for organizations.
It induces a level of cost effectiveness thereby reducing
the load for administrating numerous credentials to access
various services. Rather, there is a need to administer one
single credential for every user regardless of the number of
services they are authorized to access. Identity Access Man-
agement (IAM) system leverages this process (within one
domain) which enables user-centric authentication. How-
ever, it should not be merely deployed to replace all logins
with a single password, otherwise, this would be at the
expense of protection; if this single login is cracked, it would
then allow the intruder access to all participated services.
Therefore, some standard protocols have been developed to
secure the credential exchanging between services, such as
Security Assertion Markup Language (SAML) [46].
Securing the authentication process in the first place is still
crucial which if it is done by utilizing the aforementioned
approaches, it would yield to keeping their downsides, such
as the need to create a complex and lengthy unrepeated with
other systems password as well as the burden of memorizing
and recalling it. Additionally, SSO assumes that the autho-
rized person who has been granted access initially is the one
continues accessing the service throughout the usage session;
which is not always the case. Moreover, typical users have
other systems that are under other autonomous domains and
organizations. As a consequence, the encumbrance of cogni-
tive memory load and carrying tokens may persist.
4.2 Federated identity
To bridge the gap between separate domains and thus allevi-
ate the burden on users, federated identity management has
risen thereby extending the SSO concept from being con-
fined to a sole domain. It aims at granting access for users
of one organization to resources offered by other organiza-
tions seamlessly. To achieve this, inter-organizational trust
relationship should be established [10,48].
Thus, there is a dire need to ensure the security of these
cross-domains credentials whilst they are being communi-
cated, which in turn leads to the development and deployment
of standards, such as OpenID, WS-Federation, and Shibbo-
leth [48,49]. Whilst some of these standards (in one way
or another) act as third party federated IAM providers,
whereby an identity provider or manager coordinates the
authentication process among the member parties of the
federation which are the services providers [50], users cre-
dentials and some other information might be passed from
one service provider to another. For instance, holders of Face-
book account are able to use the credentials to access Yahoo
services although they are distinct organizations. Hence,
Facebook might send some basic information about the user,
such as name, email, mobile number and photo. Accord-
ingly, user privacy concerns must be overcome so that the
user should have the discretion to decide which of their data
can be shared, with whom and when.
Equally important, it is argued that federated identity
is fragile to breach proliferation if one of the associated
services providers’ credentials hacked. However, Madsen
et al. claimed that some of the mentioned standards offer
mechanism to contain such a breach by de-federation [50].
Nevertheless, the time scale until such containment occurs
is critical and dependent on whether it has been detected.
As a result, an efficient federated IAM system must provide
an effective auditing feature which poses issues on how to
manage it on heterogeneous domains. In addition, whereas
federated IAM approach offers promising usability advan-
tages, still replacing all passwords with a single password is
against good security practice of differing passwords for each
system. Moreover, it is still performed at the point-of-entry
leaving the system at risk of misuse afterwards. Furthermore,
Cluster Comput
it focusses upon system/service level authentication—rather
than actually looking at what the user is doing.
5 Continuous and transparent authentication
Further consideration has been given to continuous and
transparent authentication in order to solve the point-of-
entry issue. It seeks to verify whether the user is gen-
uine in a periodic or constant manner utilizing biometrics
without interrupting the user’s normal interaction [7,51].
Transparent Authentication Systems (TAS) have been stud-
ied by several researchers with varying approaches. After
a thorough analysis of the related literature, a number
of relevant search keywords have been identified within
user authentication domain, i.e. “transparent”, “continu-
ous”, “implicit”, “active”, “passive”, “non-intrusive”, “non-
observable”, “adaptive”, “unobtrusive”, and “progressive”
from various eminent academic databases. Accordingly, 93
studies have been reviewed, most of which (70 %) only
employ single biometric, as demonstrated in Fig. 2.
As each of these models (shown in Table 2) utilizes a
sole modality, they continue in carrying its shortcomings,
thus enduring low matching performance, limited universal-
ity and higher vulnerability to spoofing attacks. Fusing more
than one biometric (multimodal) can arguably contribute to
overcoming or at least alleviating these flaws [117119].
5.1 Multimodal authentication systems
Based upon analyzing the prior art on continuous and
transparent multimodal authentication systems, the 28 stud-
ies are categorized into: physiological multimodal systems;
behavioral multimodal systems; hybrid multimodal systems;
distributed multimodal systems; and web- and cloud-based
multimodal systems. The first three categories are according
to the nature of the utilized biometric modalities, whereas the
last two ones are according to their operational deployments
that distinguish them from the others.
5.1.1 Physiological multimodal systems
Table 3demonstrates proposed frameworks in this domain
deployed a set of two traits from face and fingerprint. [120]
consolidated facial and fingerprint recognition systems and
integrated their resultant output with the lapsed time to act
In evaluating their work, they proposed and used new per-
formance measures, namely: Time to Correct Reject (TCR),
Tabl e 2 Single biometric transparent authentication systems
Modality Refs.
Behavioral Keystroke [5266]
Mouse [6777]
Signature [78]
Gait [7987]
Vo i c e [ 8891]
Behavioral profiling [9296]
Physiological Face [97101]
Ear [102105]
Finger [106,107]
Palmprint [108]
Iris [109116]
Fig. 2 Continuous and
transparent authentication
Cluster Comput
Tabl e 3 Physiological Transparent Multimodal Systems
Ref. Platform Biometrics Performance (%) Experiment
Mode Limitations Features
F FP I Match FAR FRR Verification
[120]PC √√ 11 participants
30 min
Real New performance
Holistic fusion
[121]PC √√ 48.6–72.5 300 min Simulation Intrusive login
[124]PC √√ 1.0 90 participants Prototype 26–42 % added
[122]PC √√ 40 participants Simulation Intrusive login
security API
[123] PC/Laptop √√ 3.0 84–97 61 participants
Real Intrusive login (I)
Fface, FP fingerprint, Iiris
Probability of Time to Correct Reject (PTCR), Usability,
and Usability-Security Characteristic Curve (USC). How-
ever, it was undertaken with only 11 users without, even, any
Whilst [121] incorporated facial recognition and finger-
print in their model, the latter was applied intrusively when
the confidence level went below the specified threshold, mak-
ing it eventually unimodal. The accomplished matching score
of 48.6–72.5 % indicates undesirable performance especially
with critical applications. Furthermore, it and the study of
[122] were merely simulation.
Similarly, [123] investigated utilizing the face and iris
modalities but again the latter in prompted in an intrusive
manner and just at the entry point. With 61 participants, the
verification rate was between 84 and 97% with a FAR was
The desirable performance of [124] (FRR of 1.0 % with 90
users) notwithstanding, they introduced an extra processing
overhead of 26–42%, raising usability (e.g. longer waiting
time) and economic issues (e.g. power consumption).
5.1.2 Behavioral multimodal systems
The hindrance of transparently employing physiological
biometrics has been evident; thus a shift to behavioral
counterparts was sought (as shown in Table 4). [125127]
proposed the utilization of keystroke and mouse dynamics
for this purpose. The last two studies were complemented
by the inputs of the touch screen (touchalytics), albeit they
achieved higher error rates (14.47 and 2.24 FAR and 1.78 and
2.10 FRR) compared to the first study (0.651 FAR and 1.312
FRR). They were, also, conducted under controlled environ-
ment with pre-specified tasks. Therefore, generalizing their
results is questionable.
Another proposal used keystroke analysis whereas com-
bining it with voice recognition of mobile phones was
presented by [7,130]. Unlike the previous frameworks, this
experiment was with a blend of real and simulated data
and achieved a keystroke EER of 10 % and a voice EER of
25 % (without an overall performance). Despite the attained
67 % reduction of intrusive authentication, the recovery was
designed to be secret PIN, hence carrying the weaknesses
mentioned in Sect. 2.1.
Vildjiounaite et al. [128] fused voice and gait recognition
within a mobile devices context and investigated its feasi-
bility offline on the usage of 31 participants. In addition to
being offline, the resultant performance occurred on a dif-
fering range from 2.0 to 12.0 % EER, making it difficult to
reflect on how it is in the actual live use.
The focus was then shifted to deploying various aspects
of behavioral profiling as in [129,131]. The former study
accomplished an EER of 5.4, 2.2 and 13.5 % when utilized
the usage of calling, text messaging, and general applica-
tions respectively with an overall of 7.03 % EER. Likewise,
the latter experiment consolidated texting linguistic profil-
ing, keystroke dynamics and behavior profiling and obtained
an EER of 12.8, 20.8 and 9.2 % respectively with an over-
all of 3.3% and a 91 % decline in the explicit authentication
requests. However, these two studies were conducted entirely
or partly on old (2004/2005) and varying offline datasets
which were joined assuming they are of the same group of
All the aforementioned frameworks can only operate on
a distinct device (a mobile or PC). Given that users nowa-
days use typically at least one from each platform, extra care
should be taken to their applicability and universality.
5.1.3 Hybrid multimodal systems
Researchers have recognized the operational complications
of installing physiological biometrics only together with the
instability of behavioral biometrics only in a continuous and
Cluster Comput
Tabl e 4 Behavioral transparent multimodal systems
Ref. Platform Biometrics Performance (%) Experiment
Mode Limitations Features
[125]PC √√ 0.651 1.312 22 participants 9
Real IDS client–server
[128] Mobile √√ 2-12 31 participants Offline
[126]PC √√ √14.47 1.78 61 participants
10 days
Real Detection time
2.20 min
[129] Mobile 7.03 76 participants Simulation Off-line dataset Analyzed
texting & apps
[7,130] Mobile √√ (K) 10 30 participants Real & 67 % reduction of
(V) 25 7 tasks Simulation
[131] Mobile 9.2 30 participants Simulation Off-line dataset
& Real
91 % reduction of
[127]PC √√ √2.24 2.10 31 participants 3
Vvoice, Mis mouse, Kkeystroke, Bbehavioral profiling, Ggait, Ttouchalytics
transparent fashion. Therefore, various studies have been
proposed deploying a mixture of physiological and behav-
ioral or soft biometrics (e.g. color of face), as summarized in
Table 5. The study of [132] was one of the initial endeavors
which aimed to operate on and protect a flight deck. Despite
the offered level of flexibility in terms of where the veri-
fication processing carried out (on-board or distributed), it
was only conceptual with no implementation, the same as
Altinok and Turk [136] investigated the plausibility of
deploying voice verification, facial recognition, and finger-
print in a multimodal continuous authentication framework.
It integrated them with the time at which they were acquired.
The consequence of this integration would create a trust level
on the user which fluctuates based upon the interval from the
last successfully captured modalities samples. Accordingly,
it and [137], alike, produced virtual data but they did not
reveal any performance results. The latter, also, endured the
problems of secret-knowledge approach as it utilized intru-
sive login using secret code. On the other hand, although the
work of [140] was simulation also, they published results
of fusing face and voice modalities of 30 simulated partici-
pants for 3 separate sessions. They accomplished a face EER
of 0.449 %, a voice EER of 0.003 % and an overall EER of
0.087 %.
Clarke et al. [139] conducted one of the most compre-
hensive experiments in this domain. They proposed a mobile
Non-Intrusive and Continuous Authentication (NICA) using
those biometric techniques existing on the device to oper-
ate in both standalone and client–server modes—achieving
favorable performance of 0.01% EER of 27 users with 60
biometric samples collected. Nevertheless, they loosened
the threshold because they utilized in-house biometric algo-
rithms which, in turn, perhaps affected the credibility of the
result. An interesting feature of NICA is that it was designed
to use the confidence level on the legitimate user (proposed
earlier) in order to align it with the user privileges to access
services that have varying risk levels.
Other studies investigated composite authentication sys-
tems of physiological and soft biometrics [141143]on
laptops. The first study experimented the fusion of the face
trait along with its soft features, such as color, and claimed to
subsequently succeed to have no FAR and an FRR of 4.17 %.
Similarly, using the same biometrics, the last two studies
achieved a recognition score of 86.88% albeit with only 7
users. Furthermore, their experiment adopted an obtrusive
login (password or face) and merely the soft biometrics were
verified throughout, which might be affected by the surround-
ing environment, leading to convenience issues of increasing
re-authentication requests.
Leveraging the advent of wearable technologies, [138]
developed a wristband to be utilized as an initial login fin-
gerprint sensor and then to constantly measure the user skin
temperature and heart rate. However, the fingerprint was only
presented at the login stage and the performance was quite
low (matching score between 40–60%). Moreover, requir-
ing an additional wristband to access a system, inherits the
downsides of tokens.
Cluster Comput
Tabl e 5 Hybrid transparent multimodal systems
Ref. Platform Biometrics Performance (%) Experimentnewline
Mode Limitations Features
F FP V M K B G SB Match FAR FRR EER Recognition
[132] Flight Deck Conceptual No experiment 2 designs:
on-board &
[136]PC √√ √ 24 participants Virtual data Integration with
[133] Mobile 2 ×1040.4 Conceptual Intrusive login
[137]PC √√ Simulation Intrusive login
[134]PC √√ Conceptual No experiment e-Learning
[138] Wearable & Laptop √√
40–60 Prototype Intrusive login
(F) Wristband
[139] Mobile √√0.01 27 participants
45 min
Real Extendable
Standalone &
[140]PC √√ 0.087 30 participants 3
Simulation Adaptive
[141]Laptop √√
0 4.17 20 participants Real Intrusive login
[135] Mobile √√ √ Conceptual No experiment Fuzzy Crypto
[142,143]Laptop √√ 86.88 7 participants Real Swarm
Fface, FP fingerprint, Vvoice, Mmouse, Kkeystroke, Bbehavioral, Ggait, Ssoft biometrics
Cluster Comput
Tabl e 6 Distributed transparent multimodal systems
Ref. Platform Biometrics Performance (%) Experiment
Mode Limitations Features
[144]PDA √√ Conceptual No experiment One user to many
[145] Mobile & PC √√√ 9 participants Prototype 42% reduction of
[146] PDA & various devices 20 participants
14 days
Real &
Utilizes Secret &
Tok en
74 % reduction of
Fface, Vis voice, Bbehavioral profiling, PS physiological signals
Tabl e 7 Web- and cloud-based transparent multimodal systems
Ref. Platform Biometrics Performance (%) Experiment
Mode Limitations Features
[147]Web √√ (M) 22.41 24 participants 8
Real Intrusive login
Bayesian fusion
(K) 24.78
[148]Web √√ (V) 10 Prototype Intrusive login
Fface, Vvoice, Kkeystroke, Mmouse
5.1.4 Distributed multimodal systems
All the aforementioned frameworks did not consider the cur-
rent fact of a user in possession of various digital devices.
Therefore, the studies presented in Table 6have been
conducted. [144] conceptually proposed deploying physi-
ological signals (e.g. blood pressure and heart beat) and
behavioral profiling.
In the one hand, [145] prototyped a progressive authen-
tication model integrating the face, voice and behavior
profiling traits, in conjunction with proximity to pre-defined
logged-in device(s). In spite of the claimed decrease of intru-
sive verification prompts by 42%, it was investigated with 9
users only and no security measures revealed.
From the same standpoint, [146] developed their Authen-
tication Aura system utilizing what authentication techniques
exist on each device, i.e. secret, behavioral profiling, and even
personal dumb objects, such as keys. Both the authentication
status of and the user confidence level on each participating
device are communicated between each other within a close
proximity to form an overall confidence. Whilst it was carried
out on a blend of real and simulated data of 20 participants,
its focus was more on usability (74 % less explicit authenti-
cation occurrences. Additionally, further examination on the
processing overhead on each device is needed.
5.1.5 Web-based multimodal systems
Traore et al. and Ceccarelli et al. [147,148] proposed a solu-
tion to mitigate the processing burden on users’ devices and
make it occurs, instead, on a web server (Table 7). In order
to secure web services, [147] fused mouse and keystroke
dynamics for continuous identity verification following a pre-
liminary secret-knowledge login. They obtained a distinct
EER for each modality (22.41 and 24.78 respectively) and
an overall EER of 8.21. Apart from the persistence issues
of intrusive login, their framework is not compatible when
using a mobile device, with which there is no mouse inputs
and the keystroke is likely to be limited.
Ceccarelli et al. [148] proposed an Internet protocol—
Context Aware Security by Hierarchical Multilevel Architec-
ture (CASHIMA)—capable to act as a multimodal biometric
authentication system. It adopted the TAS user confidence
(trust) notion that is fluctuating based on the captured
biometrics’ time and quality, upon which users privileges
are authorized remotely. In assessing their prototype, they
integrated facial and voice recognition, on a smartphone.
Nonetheless, they did not reveal an overall performance but
for individual trait; False Match Rate (FMR) of 2.58 and 10 %
respectively. Despite the promising universality features their
framework offered, it needs to be extensively evaluated with
Cluster Comput
real data not just as a prototype to examine various metrics,
such as feasibility, scalability, and privacy-preserving.
5.2 Users’ perceptions of multimodal TAS
Crawford and Renaud and Clarke et al. [7,139] investi-
gated the users’ perceptions and acceptance of transparent
authentication. They found that 92 % of 27 participants and
73 % of 30 participants, respectively, believed that transpar-
ent authentication provided a more secure environment than
other conventional authentication. Accordingly, 90 % of the
latter’s participants stated that they would use the transparent
authentication technique if it is offered to them. The relative
small samples of both studies notwithstanding, TAS can be
appreciated as a remarkable solution to effectively remove
the reliance upon the human aspects to ensure a robust and
usable authentication. On the one hand, 83 % of 470 respon-
dents who own smartphone and tablet would like to have
seamless experience across all their devices [149].
5.3 Open issues on previous studies
As the research revolving transparent authentication evolves,
so do its evaluation and feasibility studies. It is apparent that a
multimodal TAS approach outweighs its single modal coun-
terpart due to proven security performance enhancement.
However, the abovementioned reviewed studies suffer from
one or more of the following open issues that need to be
tackled in future research:
5.3.1 Lack of transparency
It is found that a few frameworks are not, operationally, fully
transparent as they integrated a form of intrusive login (i.e.
secret, fingerprint or iris). This leads them to carry the limita-
tions of secret-knowledge authentication approach, the single
modality, and intrusive authentication.
5.3.2 Lack of universality
The majority of them, also, are confined to work in a specific
context and/or device, rendering them to lack the universal-
ity attribute that enables a seamless technology and service
independent functioning.
5.3.3 Negligence of varying services risk levels
Some studies consider the fluctuation of user identity con-
fidence/trust. Nevertheless, a little of them aligned it with
the varying risk levels of conducted activities or accessed
services, which is not the case with the real use.
5.3.4 Incomprehensive evaluation
In terms of evaluation, those studies showed performance
results carried out their experiments either on simulated/
semi-simulated data or real but insufficient and offline data.
In addition, some of them focused on usability solely whilst
others on security only. Moreover, there were specific tasks
for participants to perform, lending it not to give a better
insight about the system when they put in real live practice.
Furthermore, some other related features would be difficult to
measure, such as scalability, privacy, and subsequently user
5.4 Desirable characteristics for an effective multimodal
In order to offer an effective multimodal TAS, it should
go through comprehensive stages, from the design to the
appraisal, bearing in mind a number of critical factors. As
a result of the thorough survey and analysis, the following
desirable requirements are concluded in order to be in place
to overcome the aforementioned open issues:
no intrusive login,
no additional device or sensor,
flexibility to deploy mixture of biometrics,
continuous user identity confidence,
services risk levels aligned with user identity confidence,
minimal processing overhead,
high scalability,
compatible with various platforms,
real and adequate number of evaluation participants,
task-free experiment,
security measures to secure and manage biometric tem-
plates database and biometric samples in transient.
5.5 A framework for federated authentication in the
Stemming from the abovementioned desirable characteris-
tics, the authors have propounded a federated biometric
authentication framework, shifting the burden of both the
authentication processing and management responsibility
to centralized Managed Authentication Service Provider
(MASP) [150]. As shown in Fig. 3, this MASP is hosted on
the cloud and receives biometric signals from and control the
verification decision of the subscribed user’s devices. These
devices can benefit from the confidence level of each other
as they are fused on the MASP and communicated to those
participating devices within a close proximity. This accumu-
lated identity confidence status is utilized in both device and
service domain as MASP would verify the user identity con-
tinuously and transparently whilst they access services on
Cluster Comput
Fig. 3 A framework of federated authentication in the cloud
the device or online depending upon their determined risk
level. For example, had the user logged into his smartphone
using a fingerprint, they would, within specified period of
time and proximity, automatically logged into their regis-
tered laptop transparently without having to re-enter their
biometrics unless the user confidence status is below the risk
level of the requested service.
Even though this model is deemed to offer a potential
solution for many issues of the aforementioned systems, it
is still solely conceptual. Therefore, it lacks required tests to
appraise acute issues, such as scalability, biometrics manage-
ment, and battery consumption of portable digital devices.
Thus, developing this proposed model and evaluating it with
real and live data will perhaps give better insight about its
feasibility and value in solving the technology and research
5.5.1 Limitations and future challenges
Despite the fact that such a model would have the provisions
of effective security and usability, it raises a number of limi-
tations as future challenges that need to be addressed in order
for it to function effectively.
Trust users and organizations are required to have a high
level of trust in a third-party authentication provider;
Scalability and response time the time spent to make an
authentication decision through the network may intro-
duce a potential delay in transit and bottle-neck at the
Privacy From an end-user perspective, preserving their
privacy thereby securing their biometrics information
(during the transfer, processing and storage) is essential.
Therefore, MASP architecture must be sensibly designed
to ensure this and eliminate misuse.
6 Conclusion
Verifying the authenticity of a user to use a digital device
or service has become crucial. Individuals, businesses and
governments undertake an ever-growing range of activities
online and via mobile devices. Unfortunately these activi-
ties, services and information are the targets of cybercrimes.
Authentication is at the vanguard of ensuring that only the
authorized user is given access; however, it has historically
endured a range of issues related to the security and usability
of the approaches. Further to this, they are still mostly func-
tioning at the point of entry, and even those performing sort
of re-authentication executing it in an intrusive manner.
The majority of frameworks that were proposed to solve
this issue deployed a single biometric to re-verify the user
in a continuous but implicit fashion. Nonetheless, they have
inherited the downsides of the utilized modality so they have
issues regarding the universality and circumvention.
Therefore, a serious move towards employing two or more
biometric modality in TAS has been taken. However, most of
the previous studies in this domain fall short in one or more
drawbacks in relation to lack of full transparency, universal-
ity, interoperability, scalability, high performance, and real
data. In order to provide users with adequate protection and
convenience, innovative robust authentication mechanisms
have to be utilized in a universal level, so they operate in a
transparent, continuous and user-friendly fashion.
Cluster Comput
1. Conrad, E., Misenar, S., Feldman, J.: Cissp Study Guide. Elsevier
Inc., Burlington (2012)
2. Wood, H.M.: The use of passwords for controlling access to
remote computer systems and services. In: Proceedings of the
June 13–16, 1977, National Computer Conference (AFIPS ’77),
pp. 27–33. ACM Press, New York, NY, USA (1977)
3. Zekri, L., Furnell, S.: Authentication based upon secret knowledge
and its resilience to impostors. Adv. Netw. Commun. Eng. 3, 30–
38 (2006)
4. O’Gorman, L.: Comparing passwords, tokens, and biometrics for
user authentication. Proc. IEEE. 91, 2021–2040 (2003)
5. Kurkovsky, S., Syta, E.: Digital natives and mobile phones: a
survey of practices and attitudes about privacy and security. In:
2010 IEEE International Symposium on Technology and Society,
pp. 441–449. IEEE (2010)
6. Symes, J.E., Clarke, N.L.: Security on mobile devices: a survey
of users’ attitudes and opinions. Adv. Commun. Comput. Netw.
Secur. 9, 59–68 (2012)
7. Crawford, H., Renaud, K.: Understanding user perceptions of
transparent authentication on a mobile device. J. Trust Manag.
1, 1–28 (2014)
8. CSID: Consumer survey: password habits, a study among Ameri-
can consumers. uploads/2012/
09/CS_PasswordSurvey_FullReport_FINAL.pdf. Accessed 18
June 2013
9. Shankdhar, P.: 10 most popular password cracking tools. http://
tools/. Accessed 30 Dec 2014
10. Clarke, N.: Transparent User Authentication: Biometrics, RFID
and Behavioural Profiling. Springer, London (2011)
11. Nelson, D., Reed, V., Walling, J.: Pictorial superiority effect. J.
Exp. Psychol. Hum. Learn. Mem. 2, 523–528 (1976)
12. Weiss, R., Luca, A. De: PassShapes: utilizing stroke based authen-
tication to increase password memorability. In: Proceedings of the
5th Nordic Conference on n Human–Computer Interaction, pp.
18–22 (2008)
13. De Luca, A., Hang, A., Brudy, F., Lindner, C., Hussmann, H.:
Touch me once and i know it’s you!: implicit authentication based
on touch screen patterns. In: The SIGCHI Conference on Human
Factors in Computing Systems, CHI 2012, pp. 987–996. Austin,
TX, USA (2012)
14. Aviv, A.J., Gibson, K., Mossop, E., Blaze, M., Smith, J.M.:
Smudge attacks on smartphone touch screens. In: Proceed-
ings of the 4th USENIX Conference on Offensive technologies,
WOOT’10 (2010)
15. Wiedenbeck, S., Waters, J., Birget, J.-C., Brodskiy, A., Memon,
N.: Authentication using graphical passwords: effects of tolerance
and image choice. In: Symposium on Usable Privacy and Security
(SOUPS) 2005 (2005)
16. van Oorschot, P., Thorpe, J.: Exploiting predictability in click-
based graphical passwords. J. Comput. Secur. 19, 669–702 (2011)
17. Charrau, D., Furnell, S., Dowland, P.: PassImages: an alternative
method of user authentication. In: Proceedings of the 4th Annual
ISOneWorldConference and Convention. Las Vegas, USA (2005)
18. English, R., Poet, R.: Towards a metric for recognition-based
graphical password security. In: 5th International Conference on
Network and System Security (NSS), 2011, pp. 6–8 (2011)
19. Microsoft: features of Windows 8.1—Microsoft Windows.
personalize=startscreen. Accessed 08 Nov 2014
20. Passfaces: passfaces personal version 1.0. http://www.passfaces.
com/personal/support/ helpmanual.htm. Accessed 05 May 2014
21. Ellis, H., Shepherd, J., Davies, G.: Identification of familiar and
unfamiliar faces from internal and external features: some impli-
cations for theories of face recognition. Perception 8, 431–439
22. Aloul, F., Zahidi, S., El-Hajj, W.: Two factor authentication using
mobile phones. In: 2009 IEEE/ACS International Conference on
Computer Systems and Applications, pp. 641–644. IEEE (2009)
23. HSBC Bank plc: Secure key: two-factor authentication |HSBC
banking-security/secure-key. Accessed 05 Nov 2014
24. Google: install google authenticator.
accounts/answer/1066447?hl=en. Accessed 05 Nov 2014
25. Furnell, S.M., Katsikas, S., Lopez, J., Patel, A.: Securing Infor-
mation and Communications Systems: Principles, Technologies,
and Applications. Artech House, Norwood (2008)
26. BBC: security firm RSA offers to replace SecurID tokens. http:// Accessed 05 May
27. Nanavati, S., Thieme, M., Nanavati, R.: Biometrics: Identity Ver-
ification in a Networked World. Wiley, New York (2002)
28. Jain, A.K., Flynn, P., Ross, A.A.: Handbook of Biometrics.
Springer, New York (2008)
29. Schouten, B., Jacobs, B.: Biometrics and their use in e-passports.
Image Vis. Comput. 27, 305–312 (2009)
30. Clarke, N., Furnell, S.: Biometrics—the promise versus the prac-
tice. Comput. Fraud Secur. 9, 12–16 (2005)
31. Goode Intelligence: Mobile Phone Biometric Security—Analysis
and Forecasts 2011–2015.
andforecasts-20112015. Accessed 08 Jan 2015
32. FBI: next generation identification.
cjis/fingerprints_biometrics/ngi. Accessed 04 June 2014
33. National Science and Technology Council: The National
Biometrics Challenge 2011.
BiometricsChallenge2011_protected.pdf. Accessed 03 June 2014
34. European Central Bank: Recommendations for the Security of
Internet Payments—Final Version After Public Consultation, Ger-
many (2013)
35. Federal Financial Institutions Examination Council: Authentica-
tion in an Internet Banking Environment. http://digitallibrary.kcci. 701. Accessed 08 May 2014
36. NatWest: NatWest personal banking |online banking. http://
online/card-reader.ashx. Accessed 08 Nov 2014
37. Lloyds Bank: Lloyds Bank—internet banking—how to log on—
help logging on.
logging-on.asp? Accessed 08 Nov 2014
38. Samba financial group: SambaOnline banking—ways to bank. banking/ways-to- bank/
samba-online.html. Accessed 08 Nov 2014
39. White, C.: Windows 8.1 will focus on biometrics for authen-
on-biometrics- for-authentication. Accessed 24 March 2014
40. O’Boyle, B.: How does the Samsung Galaxy S5 fingerprint scan-
ner work? how-does-
the-samsung- galaxy-s5-fingerprint-scanner-work. Accessed 13
June 2014
41. Samsung: Samsung Galaxy S5 (Black)—review, specs
& features—Samsung UK.
uk/consumer/mobile- devices/smartphones/android/SM-
G900FZKABTU. Accessed 08 Nov 2014
42. Mogull, R.: The iPhone 5s fingerprint reader: what you need to
5s-fingerprint- reader-what-you- need-to- know.html. Accessed
13 June 2014
Cluster Comput
43. Apple: iPhone 5s—technical specifications.
com/uk/iphone- 5s/specs/. Accessed 08 Nov 2014
44. IDC: smartphone OS market share 2014, 2013, 2012,
and 2011.
share.jsp. Accessed 08 Jan 2015
45. Macworld: take the pain out of two-factor authentication with an
app. the-pain-
out-of- two-factor-authentication- with-an-app.html. Accessed
08 Jan 2015
46. Sandhu, S.: Single sign on concepts & protocols. https://www. whitepapers/authentication/single- sign-
concepts-protocols- 1352. Accessed 26 Mar 2014
47. Furnell, S.: Authenticating ourselves: will we ever escape the pass-
word? Netw. Secur. 2005, 8–13 (2005)
48. Stihler, M., Santin, A.O., Marcon Jr., A.L., Fraga, J.D.S.: Inte-
gral federated identity management for cloud computing. In: 2012
5th International Conference on New Technologies, Mobility and
Security (NTMS), pp. 1–5. IEEE (2012)
49. Cloud Security Alliance: Identity and access management imple-
mentation guidance. In: Cloud Security Alliance Security as a
Service Implementation Guidance Version 1.0. pp. 1–43 (2012)
50. Madsen, P., Koga, Y., Takahashi, K.: Federated identity manage-
ment for protecting users from ID theft. In: Proceedings of the
2005 Workshop on Digital identity management—DIM ’05, pp.
77–83. ACM Press, New York, NY, USA (2005)
51. Traore, I., Ahmed, A.A.E.: Continuous Authentication Using Bio-
metrics: Data, Models, and Metrics. IGI Global, Hershey (2012)
52. Umphress, D., Williams, G.: Identity verification through key-
board characteristics. Int. J. Man. Mach. Stud. 23, 263–273 (1985)
53. Leggett, J., Williams, G.: Verifying identity via keystroke charac-
teristics. Int. J. Man. Mach. Stud. 28, 67–76 (1988)
54. Shepherd, S.: Continuous authentication by analysis of keyboard
typing characteristics. In: European Convention on Security and
Detection, 1995, pp. 111–114. IET, Brighton (1995)
55. Mahar, D., Napier, R., Wagner, M., Laverty, W., Henderson, R.,
Hiron, M.: Optimizing digraph-latency based biometric typist ver-
ification systems: inter and intra typist differences in digraph
latency distributions. Int. J. Hum. Comput. Stud. 43, 579–592
56. Furnell, S.M., Morrissey, J.P., Sanders, P.W., Stockel, C.T.: Appli-
cations of keystroke analysis for improved login security and
continuous user authentication. Information Systems Security, pp.
283–294. Chapman & Hall, Ltd., London (1996)
57. Monrose, F., Rubin, A.D.: Keystroke dynamics as a biometric for
authentication. Future Gener. Comput. Syst. 16, 351–359 (2000)
58. Dowland, P.S., Singh, H., Furnell, S.M.: A preliminary investiga-
tion of user authentication using continuous keystroke analysis.
In: 8th IFIP Annual Working Conference on Information Security
Management and Small System Security (2001)
59. Bergadano, F., Gunetti, D., Picardi, C.: User authentication
through keystroke dynamics. ACMTrans. Inf. Syst. Secur. 5, 367–
397 (2002)
60. Gunetti, D., Picardi, C.: Keystroke analysis of free text. ACM
Trans. Inf. Syst. Secur. 8, 312–347 (2005)
61. Hempstalk, K.: Continuous typist verification using machine
learning, PhD Thesis, The University of Waikato (2009)
62. Hossain, M., Balagani, K.S., Phoha, V.V.: New impostor score
based rejection methods for continuous keystroke verification
with weak templates. In: 2012 IEEE Fifth International Confer-
ence on Biometrics: Theory, Applications and Systems (BTAS),
pp. 251–258 (2012)
63. Marsters, J.: Keystroke dynamics as a biometric. University of
Southampton (2009)
64. Messerman, A., Mustafic, T., Camtepe, S.A., Albayrak, S.:
Continuous and non-intrusive identity verification in real-time
environments based on free-text keystroke dynamics. In: 2011
International Joint Conference on Biometrics Compendium, IEEE
Biometrics (IJCB), pp. 1–8. IEEE (2011)
65. Obaidat, M.S., Sadoun, B.: Verification of computer users using
keystroke dynamics. IEEE Trans. Syst. Man. Cybern. B. Cybern.
27, 261–269 (1997)
66. Roth, J., Liu, X., Metaxas, D.: On continuous user authentication
via typing behavior. IEEE Trans. IMAGE Process. 23, 4611–4624
67. Gamboa, H., Fred, A.: A behavioral biometric system based on
human–computer interaction. In: Defense and Security, pp. 381–
392. International Society for Optics and Photonics (2004)
68. Pusara, M., Brodley, C.E.: User re-authentication via mouse
movements. In: Proceedings of the 2004 ACM workshop
on Visualization and data mining for computer security—
VizSEC/DMSEC ’04, pp. 1–8. ACM Press, New York, NY, USA
69. Ahmed, A.A.E., Traore, I.: A new biometric technology based
on mouse dynamics. IEEE Trans. dependable Secur. Comput. 4,
165–179 (2007)
70. Aksari, Y., Artuner, H.: Active authentication by mouse move-
ments. In: ISCIS 2009, 24th International Symposium on Com-
puter and Information Sciences, 2009, pp. 571–574. IEEE (2009)
71. Shen, C., Cai, Z., Guan, X., Huilan, I., Du, J.: Feature analysis
of mouse dynamics in identity authentication and monitoring. In:
IEEE International Conference on Communications, 2009, ICC
’09, pp. 1–5 (2009)
72. Zheng, N., Paloski, A., Wang, H.: An efficient user verification
system via mouse movements. In: Proceedings of the 18th ACM
Conference on Computer and Communications Security, pp. 139–
150. ACM, New York, NY, USA (2011)
73. Jorgensen, Z., Yu, T.: On mouse dynamics as a behavioral biomet-
ric for authentication. In: Proceedings of the 6th ACM Symposium
on Information. Computer and Communications Security, pp.
476–482. ACM, New York, NY, USA (2011)
74. Lin, C., Chang, C., Liang, D.: A new non-intrusive authentication
approach for data protection based on mouse dynamics. In: 2012
International Symposium on Biometrics and Security Technolo-
gies, pp. 9–14. IEEE (2012)
75. Feher, C., Elovici, Y., Moskovitch, R., Rokach, L., Schclar, A.:
User identity verification via mouse dynamics. Inf. Sci. (NY) 201,
19–36 (2012)
76. Mondal, S., Bours, P.: Continuous authentication using mouse
dynamics. In: 2013 International Conference of the Biometrics
Special Interest Group (BIOSIG), pp. 1–12. IEEE (2013)
77. Stanic, M.: Continuous user verification based on behavioral bio-
metrics using mouse dynamics. In: Proceedings of the ITI 2013
35th International Conference on Information Technology Inter-
faces, pp. 251–256. IEEE, Cavtat, Croatia (2013)
78. Clarke, N.L., Mekala, A.R.: The application of signature recog-
nition to transparent handwriting verification for mobile devices.
Inf. Manag. Comput. Secur. 15, 214–225 (2007)
79. Kale, A., Rajagopalan, A.N., Cuntoor, N., Kruger, V.: Gait-based
recognition of humans using continuous HMMs. In: Proceedings
of the Fifth IEEE International Conference on Automatic Face
and Gesture Recognition (FGRí02), pp. 1–6. IEEE (2002)
80. Morris, S.: A shoe-integrated sensor system for wireless gait
analysis and real-time therapeutic feedback. University of
Southampton (2004)
81. Mäntyjärvi, J., Lindholm, M., Vildjiounaite, E., Mäkelä, S.-M.,
Ailisto, H.: Identifying users of portable devices from gait pattern
with accelerometers. In: Proceedings. (ICASSP ’05). IEEE Inter-
national Conference on Acoustics, Speech, and Signal Processing,
2005, pp. 973–976. IEEE (2005)
82. Gafurov, D., Snekkenes, E.: Gait recognition using wearable
motion recording sensors. EURASIP J. Adv.Signal Process. 2009,
415817 (2009)
Cluster Comput
83. Derawi, M.O., Gafurov, D., Bours, P.: Towardscontinuous authen-
tication based on gait using wearable motion recording sensors.
In: Traore, I., Ahmed, A.A.E. (eds.) Continuous Authentication
Using Biometrics: Data, Models, and Metrics, pp. 170–190. IGI
Global, Hershey (2012)
84. Juefei-Xu, F., Bhagavatula, C., Jaech, A., Prasad, U., Savvides,
M.: Gait-id on the move: pace independent human identifica-
tion using cell phone accelerometer dynamics. In: 2012 IEEE
Fifth International Conference on Biometrics: Theory. Applica-
tions and Systems (BTAS), pp. 8–15. IEEE, Arlington, VA (2012)
85. Nickel, C., Wirtl, T., Busch, C.: Authentication of smartphone
users based on the way they walk using k-NN algorithm. In: 2012
Eighth International Conference on Intelligent Information Hid-
ing and Multimedia Signal Processing, pp. 16–20. IEEE (2012)
86. Lu, H., Huang, J., Saha, T., Nachman, L.: Unobtrusive gait ver-
ification for mobile phones. In: Proceedings of the 2014 ACM
International Symposium on Wearable Computers—ISWC ’14.
pp. 91–98. ACM Press, New York, NY, USA (2014)
87. Tanviruzzaman, M., Ahamed, S.I.: Your phone knows you: almost
transparent authentication for smartphones. In: 2014 IEEE 38th
Annual Computer Software and Applications Conference, pp.
374–383. IEEE (2014)
88. Woo, R.H., Park, A., Hazen, T.J.: The MIT mobile device speaker
verification corpus: data collection and preliminary experiments.
In: IEEE Odyssey 2006: The Speaker and Language Recognition
Workshop, 2006, pp. 1–6. IEEE (2006)
89. Kunz, M., Kasper, K., Reininger, H., Möbius, M., Ohms, J.: Con-
tinuous speaker verification in realtime. In: Proceedings of the
Special Interest Group on Biometrics and Electronic Signatures,
BIOSIG, vol. 2011, pp. 79–88 (2011)
90. Martucci, L.A., Zuccato, A., Smeets, B., Habib, S.M., Johansson,
T., Shahmehri, N.: Privacy, security and trust in cloud computing:
the perspective of the telecommunication industry. In: 9th Interna-
tional Conference on Ubiquitous Intelligence and Computing and
9th International Conference on Autonomic and Trusted Comput-
ing (UIC/ATC), 2012, pp. 627–632. IEEE (2012)
91. Abdullah, M., Bashier, H., Sayeed, S., Yusof, I., Azman, A.,
Ibrahim, S.Z., Liew, T.H.: Answering incoming call for implicit
authentication using smartphone. J. Theor. Appl. Inf. Technol. 61,
193–199 (2014)
92. Aupy, A., Clarke, N.: User authentication by service utilisation
profiling. In: Proceedings of the ISOneWorld 2005. Las Vegas,
USA (2005)
93. Yazji, S., Chen, X., Dick, R.P., Scheuermann, P.: Implicit user
re-authentication for mobile devices. Ubiquitous Intelligence and
Computing, pp. 1–15. Springer, New York (2009)
94. Jakobsson, M., Shi, E., Golle, P., Chow, R.: Implicit authentica-
tion for mobile devices. In: The 4th USENIX Conference on Hot
Topics in Security, HotSec’09 (2009)
95. Saevanee, H., Clarke, N., Furnell, S.: SMS linguistic profiling
authentication on mobile device. In: 2011 5th International Con-
ference on Network and System Security, pp. 224–228. IEEE
96. Li, F., Wheeler, R., Clarke, N.: An evaluation of behavioural pro-
filing on mobile devices. In: Proceedings of Second International
Conference on HAS, vol. 8533, pp. 330–339 (2014)
97. Klosterman, A., Ganger, G.: Secure continuous biometric-
enhanced authentication. In: Technical Report CMU-CS-00-134,
Carnegie Mellon University (2000)
98. Liu, X., Chen, T.: Video-based face recognition using adap-
tive hidden markov models. In: Proceedings of the 2003 IEEE
Computer Society Conference on Computer Vision and Pattern
Recognition (CVPR’03). IEEE (2003)
99. Janakiraman, R., Kumar, S., Sim, T.: Using continuous
face verification to improve desktop security. In: 2005 Sev-
enth IEEE Workshops on Applications of Computer Vision
(WACV/MOTION’05), vol. 1, pp. 501–507. IEEE (2005)
100. Clarke, N., Karatzouni, S., Furnell, S.: Transparent facial recog-
nition for mobile devices. In: Proceedings of the 7th Security
Conference. Las Vegas, USA (2008)
101. Xiao, Q., Yang, X.-D.: Facial recognition in uncontrolled condi-
tions for information security. EURASIP J. Adv. Signal Process.
2010, 1–10 (2010)
102. Hurley, D., Nixon, M., Carter, J.: Automatic ear recognition by
force field transformations. In: IEE Colloquium on Visual Bio-
metrics, pp. 2–6. IET, London (2000)
103. Rodwell, P.M.: Non-intrusive subscriber authentication for next
generation mobile communication systems, PhD Thesis, Univer-
sity of Plymouth (2006)
104. Islam, S., Davies, R., Mian, A.S., Bennamoun, M.: A fast and
fully automatic ear recognition approach based on 3D local sur-
face features. Advanced Concepts for Intelligent Vision Systems.
Lecture Notes in Computer Science, vol. 5259, pp. 1081–1092.
Springer, Berlin (2008)
105. Fahmi, P.N.A., Kodirov, E., Choi, D.-J., Lee, G.-S., Mohd Fikri
Azli, A., Sayeed, S.: Implicit authentication based on ear shape
biometrics using smartphone camera during a call. In: 2012
IEEE International Conference on Systems, Man, and Cybernetics
(SMC), pp. 2272–2276. IEEE (2012)
106. Feng, T., Liu, Z., Kwon, K.-A., Shi, W., Carbunar, B., Jiang, Y.,
Nguyen, N.: Continuous mobile authentication using touchscreen
gestures. In: 2012 IEEE Conference on Technologies for Home-
land Security (HST), pp. 451–456. IEEE (2012)
107. Koundinya, P., Theril, S., Feng, T., Prakash, V., Bao, J., Shi, W.:
Multi resolution touch panel with built-in fingerprint sensing sup-
port. In: Design, Automation & Test in Europe Conference &
Exhibition (DATE), 2014, pp. 1–6. IEEE Conference Publica-
tions, New Jersey (2014)
108. Kisku, D.R., Gupta, P., Sing, J.K., Tistarelli, M., Hwang, C.J.:
Low level multispectral palmprint image fusion for large scale
biometrics authentication. In: Traore, I., Ahmed, A.A.E. (eds.)
Continuous Authentication Using Biometrics: Data, Models, and
Metrics, pp. 89–104. IGI Global, Hershey (2012)
109. Wildes, R.: Iris recognition: an emerging biometric technology.
Proc. IEEE. 85, 1348–1363 (1997)
110. Matey, J.R., Naroditsky, O., Hanna, K., Kolczynski, R., LoIacono,
D.J., Mangru, S., Tinker, M., Zappia, T.M., Zhao, W.Y.: Iris on the
move: acquisition of images for iris recognition in less constrained
environments. Proc. IEEE. 94, 1936–1947 (2006)
111. Proença, H., Alexandre, L.: Iris segmentation methodology for
non-cooperative recognition. IEE Proc. Vis. Image Signal Process.
153, 199–205 (2006)
112. Du, Y., Arslanturk, E., Zhou, Z., Belcher, C.: Video-based non-
cooperative iris image segmentation. IEEE Trans. Syst. Man,
Cybern. B Cybern. 41, 64–74 (2011)
113. Yang, K., Du, E.: A multi-stage approach for non-cooperative iris
recognition. In: 2011 IEEE International Conference on Systems,
Man, and Cybernetics (SMC), pp. 3386–3391. IEEE (2011)
114. Chen, R., Lin, X., Ding, T.: Liveness detection for iris recognition
using multispectral images. Pattern Recognit. Lett. 33, 1513–1519
115. Mock, K., Hoanca, B., Weaver, J., Milton, M.: Real-time con-
tinuous iris recognition for authentication using an eye tracker.
In: Proceedings of the 2012 ACM Conference on Computer and
Communications Security, pp. 1007–1009. ACM (2012)
116. Sui, Y., Zou, X., Du, E.Y., Li, F.: Secure and privacy-preserving
biometrics based active authentication. In: 2012 IEEE Interna-
tional Conference on Systems, Man, and Cybernetics (SMC), pp.
1291–1296. IEEE (2012)
Cluster Comput
117. Jain, A., Nandakumar, K., Ross, A.: Score normalization in
multimodal biometric systems. Pattern Recognit. 38, 2270–2285
118. Ross, A., Nandakumar, K., Jain, A.: Handbook of Multibiomet-
rics. Springer, New York (2006)
119. De Oliveira, A.E., Henrique Matos Bezerra Motta, G., Vidal
Batista, L.: A multibiometric access control architecture for con-
tinuous authentication. In: 2010 IEEE International Conference
on Intelligence and Security Informatics, pp. 171–171. IEEE
120. Sim, T., Zhang, S., Janakiraman, R., Kumar, S.: Continuous ver-
ification using multimodal biometrics. IEEE Trans. Pattern Anal.
Mach. Intell. 29, 687–700 (2007)
121. Azzini, A., Marrara, S.: Impostor users discovery using a
multimodal biometric continuous authentication fuzzy system.
Knowledge-Based Intell. Inf. Eng. Syst. 5178, 371–378 (2008)
122. De Oliveira, A.E., Motta, G.H.M.B.: A security API for multi-
modal multi-biometric continuous authentication. In: 2011 Sev-
enth International Conference on Computational Intelligence and
Security, pp. 988–992. IEEE (2011)
123. Tsatsoulis, P.D., Jaech, A., Batie, R., Savvides, M.: Multimodal
biometric hand-off for robust unobtrusive continuous biometric
authentication. In: Traore, I., Ahmed, A.A.E. (eds.) Continuous
Authentication Using Biometrics: Data, Models, and Metrics, pp.
68–88. IGI Global, Hershey (2012)
124. Kwang, G., Yap, R.H., Sim, T., Ramnath, R.: A usability study
of continuous biometrics authentication. In: Tistarelli, M., Nixon,
M.S. (eds.) Proceedings of the Third International Conference on
Advances in Biometrics, pp. 828–837. Springer, Berlin (2009)
125. Ahmed, A., Traore, I.: Anomaly intrusion detection based on
biometrics. In: Proceedings of the 2005 IEEE Workshop on Infor-
mation Assurance and Security, pp. 452–453. IEEE (2005)
126. Pusara, M.: An Examination of User Behavior for User Re-
authentication. ProQuest, Ann Arbor (2007)
127. Bailey, K.O., Okolica, J.S., Peterson, G.L.: User identification and
authentication using multi-modal behavioral biometrics. Comput.
Secur. 43, 77–89 (2014)
128. Vildjiounaite, E., Mäkelä, S., Lindholm, M., Riihimäki, R.:
Unobtrusive multimodal biometrics for ensuring privacy and
information security with personal devices. In: Proceedings of
the 4th International Conference on Pervasive Computing, pp.
187–201. Springer, Berlin (2006)
129. Li, F., Clarke, N., Papadaki, M., Dowland, P.: Behaviour pro-
filing for transparent authentication for mobile devices. In: the
10th European Conference on Information Warfare and Security
(ECIW 2011), pp. 307–314, Tallinn, Estonia (2011)
130. Crawford, H., Renaud, K., Storer, T.: A framework for continu-
ous, transparent mobile device authentication. Comput. Secur. 39,
127–136 (2013)
131. Saevanee, H., Clarke, N., Furnell, S., Biscione, V.: Text-based
active authentication for mobile devices. IFIP Adv. Inf. Commun.
Technol. ICT Syst. Secur. Priv. Prot. 428, 99–112 (2014)
132. Carrillo, C.: Continuous Biometric Authentication for Autho-
rized Aircraft Personnel: A Proposed Design. Naval Postgraduate
School, Monterey (2003)
133. Clarke, N., Furnell, S.: A composite user authentication architec-
ture for mobile devices. J. Inf. Warf. 5, 11–29 (2006)
134. Asha, S., Chellappan, C.: Authentication of e-learners using mul-
timodal biometric technology. In: International Symposium on
Biometrics and Security Technologies, 2008, ISBAST 2008, pp.
1–6. IEEE (2008)
135. Muaaz, M.: A transparent and continuous biometric authentica-
tion framework for user-friendly secure mobile environments. In:
The 2013 ACM International Joint Conference on Pervasive and
Ubiquitous Computing (UbiComp 2013 Adjunct), pp. 4–7. ACM,
Zurich, Switzerland (2013)
136. Altinok, A., Turk, M.: Temporal integration for continuous mul-
timodal biometrics. In: Multimodal User Authentication (2003)
137. Kang, H.-B., Ju, M.-H.: Multi-modal feature integration for secure
authentication. In: Huang, D.-S., Li, K., Irwin, G.W. (eds.) Pro-
ceedings of the 2006 International Conference on Intelligent
Computing, pp. 1191–1200. Springer, Berlin (2006)
138. Ojala, S., Keinanen, J., Skytta, J.: Wearable authentication device
for transparent login in nomadic applications environment. In:
2nd International Conference on Signals, Circuits and Systems,
pp. 1–6 (2008)
139. Clarke, N., Karatzouni, S., Furnell, S.: Flexible and transparent
user authentication for mobile devices. In: Gritzalis, D., Lopez, J.
(eds.) Emerging Challenges for Security, Privacy and Trust, 24th
IFIP TC 11 International Information Security Conference, SEC
2009, pp. 1–12. Springer, Pafos, Cyprus (2009)
140. Soltane, M., Doghmane, N., Guersi, N.: Face and speech based
multi-modal biometric authentication. Int. J. Adv. Sci. Technol.
21, 41–56 (2010)
141. Niinuma, K., Park, U., Jain, A.K.: Soft biometric traits for con-
tinuous user authentication. IEEE Trans. Inf. Forensics Secur. 5,
771–780 (2010)
142. Tsai, P., Khan, M.K., Pan, J., Liao, B.: Interactive artificial bee
colony supported passive continuous authentication system. IEEE
Syst. J. IEEE Biom. Compend. 8, 395–405 (2014)
143. Khan, M.K., Tsai, P.-W., Pan, J.-S., Liao, B.-Y.: Biometric driven
initiative system for passive continuous authentication. In: 7th
International Conference on Information Assurance and Security
(IAS), 2011, pp. 139–144. IEEE (2011)
144. Chowdhury, M., Light, J., McIver, W.: A framework for con-
tinuous authentication in ubiquitous environments. In: Sixth
International Conference on Wireless Communication and Sensor
Networks (WCSN), pp. 1–6. IEEE Press (2010)
145. Riva, O., Qin, C., Strauss, K., Lymberopoulos, D.: Progressive
authentication: deciding when to authenticate on mobile phones.
In: The 21st USENIX Security Symposium (2012)
146. Hocking, C.G., Furnell, S.M., Clarke, N.L., Reynolds, P.L.: Co-
operative user identity verification using an Authentication Aura.
Comput. Secur. 39, 486–502 (2013)
147. Traore, I., Woungang, I., Obaidat, M.S., Nakkabi, Y., Lai, I.: Com-
bining mouse and keystroke dynamics biometrics for risk-based
authentication in web environments. In: 2012 Fourth International
Conference on Digital Home, pp. 138–145. IEEE (2012)
148. Ceccarelli, A., Montecchi, L., Brancati, F., Lollini, P., Marguglio,
A., Bondavalli, A.: Continuous and transparent user identity ver-
ification for secure internet services. IEEE Trans. Dependable
Secur. Comput. 12, 270–283 (2014)
149. Salesforce: 2014 Mobile Behavior Report. https://www.
2014mobilebehaviorreport.pdf. Accessed 17 Nov 2014
150. Al Abdulwahid, A., Clarke, N., Furnell, S., Stengel, I.: A concep-
tual model for federated authentication in the cloud. In: Proceed-
ings of the 11th Australian Information Security Management
Conference (AISM2013), pp. 1–11. Edith Cowan University,
Perth, Western Australia (2013)
Cluster Comput
Abdulwahid Al Abdulwahid is
a PhD researcher in the Centre
for Security, Communications
and Network Research at Ply-
mouth University (UK), and a
lecturer in the Computer Sci-
ence and Engineering Depart-
ment at Jubail University Col-
lege (KSA). He obtained his BSc
in Computer Information Sys-
tems and MSc in Management
of Information Technology from
King Faisal University and Not-
tingham University, (in 2003 and
2010 respectively). His research
interests include user authentication, biometrics, and cloud security and
privacy. Abdulwahid has published a number of peer reviewed publi-
cations at international conferences and is a member of the Center of
Excellence in Information Assurance (KSA) and a professional mem-
ber of the ACM and its Special Interest Group on Security, Audit and
Control (ACM SIGSAC).
Prof. Nathan Clarke is a Pro-
fessor in Cyber Security and
Digital Forensics at Plymouth
University. His research interests
reside in the area of informa-
tion security, biometrics, foren-
sics and cloud security. Prof.
Clarke has over 140 outputs con-
sisting of journal papers, confer-
ence papers, books, edited books,
book chapters and patents. He is
the Chair of the IFIP TC11.12
Working Group on the Human
Aspects of Information Security
& Assurance. Prof. Clarke is a
chartered engineer, a fellow of the British Computing Society (BCS)
and a senior member of the IEEE. He is the author of Transparent
Authentication: Biometrics, RFID and Behavioural Profiling published
by Springer. Prof. Clarke is currently involved in the £1M EPSRC
funded project ‘Identifying and Modelling Victim, Business, Regula-
tory and Malware Behaviours in a Changing Cyberthreat Landscape’
and has been involved in a number of successful Knowledge Transfer
Projects and EU Framework 7 projects.
Ingo Stengel is Professor at
the Faculty of Computer Science
and Business Computer Science
of the University of Applied Sci-
ences Karlsruhe, Germany. He
has international teaching and
research experience with several
international institutions includ-
ing Plymouth University, UK
and Cork Institute of Technol-
ogy, Ireland. His research inter-
ests are in the areas of: IT Secu-
rity, Information Security Man-
agement and eBusiness. He pub-
lished and contributed to a vast
number of peer reviewed publications at international conferences and
in international journals.
Prof. Steven Furnell is the
head of the Centre for Secu-
rity, Communications & Net-
work Research at Plymouth Uni-
versity (UK), an Adjunct Pro-
fessor with Edith Cowan Uni-
versity (Western Australia) and
an Honorary Professor with Nel-
son Mandela Metropolitan Uni-
versity (South Africa). His inter-
ests include cybercrime, mobile
device security, user authentica-
tion, and security usability. Prof.
Furnell is the author of over 260
papers in refereed international
journals and conference proceedings, as well as books including Cyber-
crime: Vandalizing the Information Society (2001) and Computer
Insecurity: Risking the System (2005). He is also the editor-in-chief
of Information & Computer Security, and the co-chair of the Human
Aspects of Information Security & Assurance (HAISA) symposium
( Steve is active in a variety of professional bodies, and
is a Fellow of the BCS, a Senior Member of the IEEE, and a Board
Member of the IISP. Further details can be found at
uk/cscan, with various security podcasts available via
Christoph Reich is a professor
at the faculty of computer science
at the University of Applied Sci-
ence in Furtwangen (HFU) and
teaches in the field of network
technologies, programming, IT
management, middleware and IT
security. He has the scientific
management of the HFU Infor-
mation and Media Center, which
consists of the departments IT,
Online Systems, Learning Sys-
tems and HFU library depart-
ment. As a director of the Insti-
tute for Cloud Computing and
IT-Security (, his research focuses on
cloud computing, QoS, virtualization and IT security.
... Continuous user authentication is an unconscious process of verifying a user based on behavioral attributes [21], also called "transparent," "implicit," "non-intrusive," "nonobservable," or "unobtrusive" [22]. Gait corresponds to continuous authentication, and it is divided into vision-based, floor sensor-based, and wearable sensor-based according to the data collection method. ...
Full-text available
With the rapid development of networking and computing technology, users can easily store and interact with sensitive information on smart devices. Since smart devices are vulnerable to unauthorized access or theft, the security of personal information is becoming more important. Gait authentication is attracting attention as a continuous or unconscious biometrics method for smart devices. However, various factors, such as gait variability and sensor state by day, can degrade authentication performance. This study proposed a sensor compensation algorithm that overcomes various factors that may occur in the real world and new 2D cyclogram features to improve user authentication performance. The dataset consists of gait data from 20 people wearing wearable sensors on the wrist and thigh over 3 days. A support vector machine (SVM) model was used for the classification of gait authentication. The results showed that the proposed sensor compensation algorithm could obtain a consistent gait signal by transforming the unstable sensor coordinate system into a stable anatomical coordinate system. Also, 2D cyclogram feature sets could be used to effectively discriminate individual gait patterns. The proposed gait authentication has an accuracy of 99.63%, 94.16%, and 94.2% and an equal error rate (EER) of 0.3%, 5.84%, and 5.8% for the same session (day 1), cross session1 (day 2), and cross session2 (day 3), respectively.
... In the past ten years, mobile devices such as smartphones, tablet computers, and smartwatches have embedded more and more sensors. The smartphone-based authentication is also called "transparent, implicit, active, nonintrusive, nonobservable, adaptive, unobtrusive, and progressive" techniques [3][4][5]. ...
Full-text available
As mobile devices become more and more popular, users gain many conveniences. It has also made smartphone makers install new software and prebuilt hardware on their products, including many kinds of sensors. With improved storage and computing power, users also become accustomed to storing and interacting with personally sensitive information. Due to convenience and efficiency, mobile devices use gait authentication widely. In recent years, protecting the information security of mobile devices has become increasingly important. It has become a hot research area because smartphones are vulnerable to theft or unauthorized access. This paper proposes a novel attack model called a collusion attack. Firstly, we study the imitation attack in the general state and its results and propose and verify the feasibility of our attack. We propose a collusion attack model and train participants with quantified action specifications. The results demonstrate that our attack increases the attacker's false match rate only using an acceleration sensor in some systems sensor. Furthermore, we propose a multi-cycle defense model based on acceleration direction changes to improve the robustness of smartphone-based gait authentication methods against such attacks. Experimental results show that our defense model can significantly reduce the attacker's success rate.
... Password and PIN are cost-effective from an implementation pointof-view and can provide fast authentication (Kunda and Chishimba, 2018). However, password and PIN have several vulnerabilities (Al Abdulwahid et al., 2016). First, users tend to use a simple and easily guessed password due to the limitation of long-term memory (Meng et al., 2018b). ...
The advancement in the computational capability and storage size of a modern mobile device has evolved it into a multi-purpose smart device for individual and business needs. The increasing usage of this device has led to the need for a secure and efficient authentication mechanism. For securing mobile devices, password, PIN, and swipe patterns are commonly used for user authentication. Entry-point face and fingerprint recognition have also gained traction in the past years. However, these authentication schemes cannot authenticate a user after the initial-login session. This limitation might put the device exposed to information theft and leakage if an illegitimate user could bypass the initial-login session. Therefore, a mobile device needs a continuous authentication mechanism that can protect a user throughout the entire working session, which complements the initial-login authentication to provide more comprehensive security protection. Touch biometric is a behavioural biometric that represents the touch behaviour pattern of a user when interacting with the touchscreen of the device. Touch biometric has been proposed as a continuous authentication mechanism, where the device can collect touch biometric data transparently while a user is using the device. However, there are still plenty of challenges and obstacles in touch-based continuous mobile device authentication due to its challenges as a biometric modality. This paper provides a comprehensive overview of fundamental principles that underpin touch-based continuous mobile device authentication. Our work discusses state-of-the-art methods in touch data acquisition, behavioural feature extraction, user classification, and evaluation methods. This paper also discusses some challenges and opportunities in the current touch-based continuous mobile device authentication domain to obtain a broad research community and market acceptance.
Full-text available
With the rapid growth of wearable devices, more applications require direct communication between wearable devices. To secure the communication between wearable devices, various pairing protocols have been proposed to generate common keys for encrypting the communication. Since the wearable devices are attached to the same body, the devices can generate common keys based on the same context by utilizing onboard sensors to capture a common biometric signal such as body motion, gait, heartbeat, respiration, and EMG signals. The context-based pairing does not need prior information to generate common keys. As context-based pairing does not need any human involvement in the pairing process, the pairing also increase the usability of wearable devices. A wide range of context-based pairing approaches has been proposed with different sensors and different biometric signals. Given the increasing popularity of wearable devices and applications of wearable devices, we believe that it is necessary to have a comprehensive review and comparison on the context-based pairing approaches for future research on the pairing. In this paper, we compare context-based pairing approaches and review common techniques used in pairing based on various biometric signals.
In this article, the challenges and future prospects in biometrics modalities are discussed. First, typical biometric modalities are introduced. Next, as current challenges in biometrics, a privacy issue and spoofing issues are discussed. For resolving the spoofing issue in system-user management, continuous authentication is necessary for which new biometrics modalities such as brain waves and intrabody propagation signals are introduced. Finally, as future prospects, smell- and taste-based authentication and some new perspectives on biometrics are introduced.
To realize a secure system-user management, continuous authentication must be implemented in the system. In addition, only limited biometrics that can be measured passively are applicable for continuous authentication. However, continuous authentication is a heavy processing load for the system. In this study, possible methods for conducting a continuous authentication are examined from the viewpoint of reducing the processing load, and two types of on-demand authentication approaches are confirmed to be effective.KeywordsSystem-user managementContinuous authenticationBiometricsOn-demand authenticationProcessing load
In recent years, user identity verification techniques based on mobile touch interaction are becoming more reliable and prominent. These techniques can be integrated in many types of mobile applications and help preventing illegitimate access to information and storage done by impostors. Such techniques usually rely on binary classification algorithms building models for each user in the system, that require other users data in order to build ones model and in case the other users’ data does not exists, a model cannot be built.In addition using a method that generates a model for each user in the system pose a significant challenge when scaling up to real-world systems. The main drawback with having a great number of models resides in the fact that it is difficult to debug, analyze, update, and fine-tune each individual model. Thus, we introduce a method for generating a unified global model that can verify the identity of every user in the system. Our most fundamental challenge is to preserve the unique behavior of each user in one unified model. Having a global model that was built prematurely also has privacy preserving characteristics, when compared to methods that require other users data. The core idea of our method is a novel behavioral embedding layer that captures and embeds each user’s unique behavior and enables it to be used within global settings. We compared our method to several state-of-the-art techniques in experiments with over 9,000 users on 85 different devices. Our method achieves 0.918 AUC and 15.6% EER in efficient global settings, bypassing the second-best method by a large margin of 0.119 on AUC and 11.7% on EER.
Purpose This paper aims to address the user perspective about usability, security and use of five authentication schemes (text and graphical passwords, biometrics and hardware tokens) from a population not covered previously in the literature. Additionally, this paper explores the criteria users apply in creating their text passwords. Design/methodology/approach An online survey study was performed in spring 2019 with university students in Mexico and Bosnia and Herzegovina. A total of 197 responses were collected. Findings Fingerprint-based authentication was most frequently perceived as usable and secure. However, text passwords were the predominantly used method for unlocking computer devices. The participants preferred to apply personal criteria for creating text passwords, which, interestingly, coincided with the general password guidelines, e.g. length, combining letters and special characters. Originality/value Research on young adults’ perceptions of different authentication methods is driven by the increasing frequency and sophistication of security breaches, as well as their significant consequences. This study provided insight into the commonly used authentication methods among youth from two geographic locations, which have not been accounted for previously.
Full-text available
Continuous authentication has been proposed as a possible approach for passive and seamless user authentication, using sensor data comprising biometric, behavioral, and context-oriented characteristics. Since these are personal data being transmitted and are outside the control of the user, this approach causes privacy issues. Continuous authentication has security challenges concerning poor matching rates and susceptibility of replay attacks. The security issues are mainly poor matching rates and the problems of replay attacks. In this survey, we present an overview of continuous authentication and comprehensively discusses its different modes, and issues that these modes have related to security, privacy, and usability. A comparison of privacy-preserving approaches dealing with the privacy issues is provided, and lastly recommendations for secure, privacy-preserving, and user-friendly continuous authentication.
Full-text available
With the growth of wearable devices, which are usually constrained in computational power and user interface, this pairing has to be autonomous. Considering devices that do not have prior information about each other, a secure communication should be established by generating a shared secret key derived from a common context between the devices. Context-based pairing solutions increase the usability of wearable device pairing by eliminating any human involvement in the pairing process. This is possible by utilizing onboard sensors (with the same sensing modalities) to capture a common physical context (e.g., body motion, gait, heartbeat, respiration, and EMG signal). A wide range of approaches has been proposed to address autonomous pairing in wearable devices. This paper surveys context-based pairing in wearable devices by focusing on the signals and sensors exploited. We review the steps needed for generating a common key and provide a survey of existing techniques utilized in each step.
Full-text available
Smartphones are being used to keep sensitive data and make private transaction other than making calls and receive short messages. Thus, authentication of the smartphones becomes very crucial and important aspect. However, users feel inconvenience and difficult with current authentication methods, from password up to physical biometrics. Implicit authentication system emerged intending to improve the security and convenience of the smartphone users. One of the approaches is considering the way users answer incoming phone calls using their smartphones. We study and evaluate the voice signal from users when answering incoming phone calls. Our study shows that the voice signals capable of authenticating the smartphone users. The experiment conducted shows a very high performance with 98.9% accuracy. These findings will promisingly augment that the novel implicit and transparent authentication system based on voices of answering incoming phone calls is feasible so that authentication of smartphone's users become easier and unobtrusive.
Biometric recognition, or simply biometrics, is a rapidly evolving field with applications ranging from accessing one's computer, to gaining entry into a country. Biometric systems rely on the use of physical or behavioral traits, such as fingerprints, face, voice and hand geometry, to establish the identity of an individual. The deployment of large-scale biometric systems in both commercial (e.g., grocery stores, amusement parks, airports) and government (e.g., US-VISIT) applications, increases the public's awareness of this technology. This rapid growth also highlights the challenges associated with designing and deploying biometric systems. Indeed, the problem of biometric recognition is a grand challenge in its own right. The past five years have seen a significant growth in biometric research resulting in the development of innovative sensors, robust and efficient algorithms for feature extraction and matching, enhanced test methodologies and novel applications. These advances have resulted in robust, accurate, secure and cost effective biometric systems. The Handbook of Biometrics -- an edited volume by prominent invited researchers in biometrics -- describes the fundamentals as well as the latest advancements in the burgeoning field of biometrics. It is designed for professionals, practitioners and researchers in biometrics, pattern recognition and computer security. The Handbook of Biometrics can be used as a primary textbook for an undergraduate biometrics class. This book is also suitable as a secondary textbook or reference for advanced-level students in computer science.
This new study guide is aligned to cover all of the material included in the exam complete with recent updates. The 10 domains are covered completely and as concisely as possible with an eye to acing the exam. Includes three practice exams.
Continuous biometric authentication is a process where the installed biometric systems continuously monitor and authenticate the users. Biometric system could be an exciting application to log in to computers and in a network system. However, due to malfunctioning in high-security zones, it is necessary to prevent those loopholes that often occur in security zones. It has been seen that when a user is logged in to such systems by authenticating to the biometric system installed, he/she often takes short breaks. In the meantime some imposter may attack the network or access to the computer system until the real user is logged out. Therefore, it is necessary to monitor the log in process of the system or network by continuous authentication of users. To accomplish this work we propose in this chapter a continuous biometric authentication system using low level fusion of multispectral palm images where the fusion is performed using wavelet transformation and decomposition. Fusion of palmprint instances is performed by wavelet transform and decomposition. To capture the palm characteristics, a fused image is convolved with Gabor wavelet transform. The Gabor wavelet feature representation reflects very high dimensional space. To reduce the high dimensionality, ant colony optimization algorithm is applied to select relevant, distinctive, and reduced feature set from Gabor responses. Finally, the reduced set of features is trained with support vector machines and accomplishes user recognition tasks. For evaluation, CASIA multispectral palmprint database is used. The experimental results reveal that the system is found to be robust and encouraging while variations of classifiers are used. Also a comparative study of the proposed system with a well-known method is presented.
Conventional access control solutions rely on a single authentication to verify a user's identity but do nothing to ensure the authenticated user is indeed the same person using the system afterwards. Without continuous monitoring, unauthorized individuals have an opportunity to "hijack" or "tailgate" the original user's session. Continuous authentication attempts to remedy this security loophole. Biometrics is an attractive solution for continuous authentication as it is unobtrusive yet still highly accurate. This allows the authorized user to continue about his routine but quickly detects and blocks intruders. This chapter outlines the components of a multi-biometric based continuous authentication system. Our application employs a biometric hand-off strategy where in the first authentication step a strong biometric robustly identifies the user and then hands control to a less computationally intensive face recognition and tracking system that continuously monitors the presence of the user. Using multiple biometrics allows the system to benefit from the strengths of each modality. Since face verification accuracy degrades as more time elapses between the training stage and operation time, our proposed hand-off strategy permits continuous robust face verification with relatively simple and computationally efficient classifiers. We provide a detailed evaluation of verification performance using different pattern classification algorithms and show that the final multi-modal biometric hand-off scheme yields high verification performance.
Conference Paper
Continuously and unobtrusively identifying the phone's owner using accelerometer sensing and gait analysis has a great potential to improve user experience on the go. However, a number of challenges, including gait modeling and training data acquisition, must be addressed before unobtrusive gait verification is practical. In this paper, we describe a gait verification system for mobile phone without any assumption of body placement or device orientation. Our system uses a combination of supervised and unsupervised learning techniques to verify the user continuously and automatically learn unseen gait pattern from the user over time. We demonstrate that it is capable of recognizing the user in natural settings. We also investigated an unobtrusive training method that makes it feasible to acquire training data without explicit user annotation.
Nearly all systems conduct some kind of user authentication before granting access to the objects or services. Moreover, humans pass through authentication steps more than once in their everyday activity, e.g. for entering a house you have to possess the correct key to open the door, to use a computer you need to know its password, etc. These authentications are one-time or static which means once the user's identity is verified the authentication lasts forever. However, some high security systems require ensuring the correct identity of the user throughout the full session. This then requires verification of user identity continuously or periodically. One of the important requirements for continuous authentication is that the method should be unobtrusive and convenient in usage. If this is not satisfied the users are not going to accept continuous authentication. Therefore not all authentication methods can be suitable for continuous authentication even if they provide higher security. In this chapter we present continuous authentication using gait biometric. Gait is a person's manner of walking and gait recognition refers to the identification and verification of an individual based on gait. This chapter discusses advantages and disadvantages of gait biometrics in the context of continuous authentication. Furthermore, we present a framework for continuous authentication using gait biometrics. The proposed framework extends on traditional static (one-time) user authentication. The framework can also be applied to other biometric modalities with small modifications.
In this chapter the Authors introduce the concepts behind the mouse dynamics biometric technology, present a generic architecture of the detector used to collect and process mouse dynamics, and study the various factors used to build the user's signature. The Authors will also provide an updated survey on the researches and industrial implementations related to the technology, and study possible applications in computer security.