Face Recognition with Liveness Detection using Eye
and Mouth Movement
Avinash Kumar Singh, Piyush Joshi, G. C. Nandi
Robotics and Artiicial Intelligence Laboratory
Indian institute of Information Technology, Allahabad - 211012, UP, india
{avinashkumarsingh 1986, piyushjoshi3839}@gmail.com, gcnandi@iiita.ac.in
Abstract- The recent literature on face recognition technology
discusses the issue of face spooing which can bypass the
authentication s
y
stem by placing a photo/video/mask of the
enrolled person in front of the camera. This problem could be
minimized b
y
detecting the liveness of the person. Therefore, in
this paper, we propose a robust liveness detection scheme based
on challenge and response method. The liveness module is added
as extra la
y
er of securit
y
before the face recognition module. The
Iiveness module utilizes face macro features, especially eye and
mouth movements in order to generate random challenges and
observing the user's response on account of this. The reliabilit
y
of liveness module is tested b
y
placing diferent t
y pes of spooing
attacks with various means, like using photograph, videos, etc. In
all, ive types of attacks have been taken care of and prevented
by our system. Experimental results show that system is able to
detect the Iiveness when subj ected to all these attacks except the
eye & mouth imposter attack. This attack is able to bypass the
liveness test but it creates massive changes in face structure.
Therefore resultant unrecognized or misclassiied b
y
the face
recognition module. An experimental test conducted on 65
persons on university of Essex face database conirms that
removal of e
y
e and nose components results 75%
misclassiication.
Keywords- Face Recognition, Liveness Detection, Face Spooing,
Face Macro Features Movement, HAAR Classiier, Principal
Component Anal
y
sis.
I. INTRODUCTION
Biometric authentication is a way to authenticate persons
that usually a human does in his life. Every human being has
some physiological or behavioural characteristics like (face,
ingerprint, voice, gait, etc.), which make them unique,
thereby diferentiating them rom others. Among all these
biometric traits, we restrict ourselves to face. The history of
face recognition technology started roml960s. From the last
50 years, face recognition technology has experienced a rapid
growth in itself [10]. Various methods have been proposed by
various researchers so far, to recognize the person even in the
bad illumination, different facial expressions, orientations, and
even in partial occlusions [1]. The latest survey [2] shows that
face recognition technology is the second most used biometric
technology used by the marke/users. Face recognition is now
used in diferent areas like access controls, human robot
interaction, surveillance, etc. Recently several researchers
quantiied the integrity of the face recognition system and
found that these systems are also vulnerable to diferent kinds
of attacks. Most requently used attack is the spooing attack
[7][8][6]. Face spooing is an attack where attacker tries to
bypass the face recognition system by placing a
photo/video/mask of the enrolled persons in ront of the
camera. The problem lies within the working principle of
Face Recognition. The principal of face recognition doesn't
care about who is submitting the credentials. It only concerns
about the person is enrolled or not. Hence whatever effort we
will do for making our classiication good and effort in using
good resolution camera, nothing will help. Besides this, it
helps attacker to perform their attack more accurately. This
problem leads to the question where one can think about the
signiicance of the accuracy and eiciency of the system,
when the reliability is not assured. Liveness detection of the
user could be a way to deal with this problem.
Therefore Researchers observed the need of security
mechanism and proposed various ways to deal with this
problem. On the basis of literature we have grouped possible
solution in three main categories (1) liveness detection by
using challenge and response method (2) liveness detection by
utilizing face texture (image quality), and (3) liveness
detection by combining two or more biometrics (multi-modal
approach). In challenge and response method system throws
some challenge in terms of eyes and mouth movement which
can only be performed by real user not by photo, and analyses
there response in account of the given challenges.In this
regard most of the researchers [3][4][15] have utilized eye
blinking, while in type-II researcher's exploits texture
information (smoothness/roughness, edges, etc.) to distinguish
between real and imposter [5][17]. Multimodal approach
mostly uses speech and face as the combination to deal with
this attack [18].
These spooing techniques are going more complex day by
day right rom simple photograph to painted contact lenses
and polymeric face, ingers. Hence a list of modem
approaches to deal with all these circumstances is mentioned
by [6][7].He has suggested, in their report that for liveness
detection, maximum utilization of face macro features is the
best way; hence in this paper we utilized both the eye and
mouth movement to detect the liveness with the constraint of
challenge and response. Previous techniques discussed in the
literature are mostly based on the eye blinking, which can be
978-1-4799-3140-8/14/$31.00 ©2014 IEEE 592