International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882

Volume 3, Issue 8, November 2014

1217

FACE AND EXPRESSION RECOGNITION IN REAL TIME

BIOMETRIC SYSTEMS
K.Akila1 T.K.Revathi2 S.Thejeswi3
1

PG Student,[2,3]Assistant Professor
Affiliated to Anna University Chennai, Dept. of Computer Science and Engineering
Kongunadu College of Engineering and Technology

Abstract:
Face recognition is a natural human ability and a widely accepted identification andverificationtechnique. In
currentlawfulsituation, a lot of credibility is located on identifications made by observer.Accordingly these are supported on
human awareness which is often flawed and can lead to situationswhere identity is disputed.One of the major
challengesencountered by face recognitionlies in the difficulty of handlingarbitrary posesvariations.Multiple expression
infused face images are generated for each subjectto enhance facerecognition performance.Facial feature points are one of the
most importantclues for manycomputer vision applications such asregistration;modelbased humanface coding andfaces
normalization.Therefore, automating the withdrawal of these points would have a widerange of usage. 3D Face recognition
has been considered as a majorsolution to deal withunsolved issues of reliable 2D face recognitionin recent years, i.e. lighting
and pose variations.Such pose variations can cause extensive occlusions resulting in missing data.However,3Dtechniques are
currently limited by their high registrationand computation cost.Analyze illumination, pose and occlusion variations in
recognitions.
Keywords: biometrics, 3D face recognition, occlusion, illumination
1. INTRODUCTION:
Recognition of humans has become a substantial topic
today as the need for security applications grows
continuously.Face is the natural assertion of identity: We
show our face as proof of who we are. Due to this widely
accepted face as proof of who we are. Due to this widely
acceptedcultural convention, face is the most widely
acceptedbiometric modality. From last few decades a
biometricbasedrecognition system becoming very useful
in a varietyof applications. Though some biometric
detection systems,such asiris and fingerprints, have
previouslyattained very highlevel of correctness, but they
have a partialutilize in nonsupportivescenarios. Biometry
enables reliable and efficient identity management systems
by exploiting physical and behavioral characteristics of the
subjects which are permanent, universal and easy to
access. Face recognition is a critical and popular topic in
the areaof computer visionand image processing for its
wide applicationpotential and scientific challenge.
Unfortunately,despitethe great progress made in the field,
as a biometricfeature, the face does not remainreliable as
affected by variationsof illumination, pose, facial
expression, etc. Recognition of a face elicits an
excitingreply, and humans are expert at identifydelicate
differences in the mood or intent of others from their facial
expressions. Facial appearance provides universally

interpretable information about a persons gender,

ethnicity, age, intelligence, their emotional state and their
health. In modern legal settings, the courts often seek
positive proof of the uniqueness. Lotcredibility is located
on identifications made by eyewitnesses in court because
human perception and recollection is erroneously
presumed to be correct on every time. In the adversarial
organization of impartiality this is the one area where a
setobserver,
who
is
usuallysimplyallowed
to
narrateaccuratedata,
is
acceptable
to
communicateaviewwith no justification. Characterizing
the variability in appearance with lighting is a fundamental
problem in many areasof face modeling,computer
graphicsand computer visualization. One of the
hugedisputes of computervision is to produce systems that
can work in uncontrolled situations. To be tough,
identificationsystems should be able to work outdoors in a
lighting-insensitive manner. In computer graphics, the
challenge is to be able to efficiently create the visual
appearance of a scene under realistic,possiblychanging,
illumination.

www.ijsret.org

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 8, November 2014

1218

conditions. In last decade, many algorithms based on

linear/nonlinear methods, neural networks, wavelets, etc.
have been proposed. Nevertheless, Face Recognition
Vendor Test 2002 shown that most of these approaches
encountered problems in outdoor conditions. This lowered
their reliability compared to state of the art biometrics.
This paper provides an ''excursus'' of recent face
recognition research trends in 2D imagery and 3D model
based algorithms. To simplify comparisons across
different approaches, tables containing different collection
of parameters are provided. This paper concludes by
proposing possible future directions.

3. PRINCIPAL COMPONENT ANALYSIS:

Figure 1: System Architecture

2. RELATED WORK:
In [1] L. D. Introna and H. Nissenbaum et al.Facial
recognition technology (FRT) has emerged as an attractive
solution to address many contemporary needs for
identyand verify of identity claims. It brings together the
promise of other biometric system, which attempt to tie
identity to individualdistinctive features of the body, and
the more thenfamiliar functionality of visual surveillance
systems. This report develops a socio-political analysis
that bridges the technical and social-scientific literatures
on FRT and addresses the unique challenges and concerns
that attend its development, evaluation, and specific
operational uses, contexts, and goals. It highlights the
potentional and limitations of the technology, noting those
tasks for which it seems ready for deployment, those areas
where performance obstacles may be overcome by future
technological developments or sound operating
procedures, and still other issues which appear as
intractable. Its concern with efficacy extends to ethical
considerations.
In[2] A. F. Abate, M. Nappi, D. Riccio, and G. Sabatino et
al. Government agencies are investing a considerable
amount of resources into improving security systems as
result of recent terrorist events that dangerously exposed
flaws and weaknesses in today's safety mechanisms.
Badge or password-based authentication procedures are
too easy to hack. Biometrics represents a valid alternative
but they suffer of drawbacks as well. Iris scanning, for
example, is very reliable but too intrusive; fingerprints are
social accepted, but not applicable to non-consentient
people. On the other hand, face recognition represents a
good compromise between what's socially acceptable and
what's reliable, even when operating under controlled

Principal component analysis is appropriate when you

have obtained measures on a number of experiential
variables and wish to develop a smaller number of
artificial variables that will account for most of the
variance
in
the
experiential
variables.
The
principalcomponents might then be utilized as predictor or
criterion variables in following analysis. Principal
component investigation is a variable decreaseprocess. It
is helpful when you havefind data on a number of
variables, and consider thatthere is a littleidleness in those
variables. In this case, redundancy means that a little of
thevariables are associated with one a different,
perhapssince they are measuring the similarconstruct.
Because of this redundancy, you believe that it should be
possible to decrease theobserved variables into a smaller
number of principal components that willaccount for most
of the discrepancy in the experiential variables.Since it is a
variable decreaseprocess, principal component study is
like in a lot ofrespects to tentative factoranalysis. In detail
the steps chased when behavior a principalcomponent
analysis are nearlythe same to those chase when
conducting anprobingfactor analysis. However, there
areimportant
abstract
differences
between
the
twoprocedures, and it is significant that you donot
wronglymaintain that you are performing factoranalysis
when you are really performingprincipal component
analysis. The differences among these two procedures
areexplained ingreater feature in a later on section titled
PrincipalComponent Analysis is Not Factor Analysis.

www.ijsret.org

Figure 2: PCA analysis

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 8, November 2014

4. LINEAR DISCRIMINANT ANALYSIS

Linear
discriminant
analysis
(LDA) and
the
interrelated Fisher's linear discriminates are techniques
utilized
in statistics, pattern
detection and machine
learning to discover a linear arrangement of features which
distinguish or divides two or more classes of objects or
events. The resultantarrangementmight be used as a linear
classifier, or, additionalusually, for dimensionality
decrease previous toafterward classification.
LDA is closely related to ANOVA and regression
analysis, which also effort to articulate one needy
variable as a linear arrangement of additional features or
measurements. However,ANOVAuses definite independe
nt variables and a continuous reliant variable, whereas
discriminant investigation has uninterrupted independent
variables and a categorical dependent variable. Logistic
regression and probit regression are more related to LDA,
as they also clarify a definite variable by the values of
continuous self-determining variables. These other
techniques are preferable in applications where it is not
reasonable to assume that the independent variables are
usually distributed, which is a basicsupposition of the
LDA technique.
LDA is also closely related to principal
component analysis (PCA) and factor analysis in that they
both look for linear combinations of variables which best
clarify the data. LDA unambiguouslyefforts to model the
dissimilarityamong the classes of data. PCA on the other
hand does not obtain into account somedissimilarity in
class, and factor analysis constructs the feature
combinations supports on dissimilarity rather than
correspondence. Discriminant analysis is also dissimilar
from factor analysis in that it is not an interdependence
technique: a distinction between independent variables and
dependent variables must be made.
LDA works when the measurements made on independent
variables for each observation are constant quantities.
When dealing with definiteself-determining variables, the
correspondingmethod is discriminant correspondence
analysis.

5. LOCAL BINARY PATTERN

Local Binary Pattern (LBP) is a simple yet very efficient
texture operator which labels the pixels of an image by
thresholding the neighborhood of each pixel and considers
the result as a binary number. Appropriate to its
discriminative power and computational ease, LBP texture
operator has develop intoanacceptedmethod in a variety of
applications. It can be observe as a combinemethod to the
traditionally divergent statistical and structural models of
texture
analysis.
Conceivably
the
majoritysignificantbelongings of the LBP operator in realworld applications is its robustness to monotonic gray-

1219

scale
revolutionizefoundation,
for
instance, by
clarification variations. One moresignificantbelonging is
its computational ease, which compose it probable to
evaluate pictures in demanding real-time settings.

Figure 3: LBP analysis

In the LBP method for texture categorization, the occasion
of the LBP codes in apicture are composed into a
histogram. The categorization is then executed by
calculateeasy histogram comparisons. Allowing for a
relatedmethod for facial picture illustrationoutcome in a
loss of spatial information and therefore one should codify
the texture information while retaining also their positions.
One method to realize this objective is to utilize the LBP
texture descriptors to build several local descriptions of
the face and combine them into a global explanation. Such
local descriptions have been aheadattention lately which is
understandable given the limitations of the holistic
illustrations. These local feature supportedtechniques are
extra robust against variations in pose or illumination than
holistic methods.
This histogram efficiently has anexplanation of the face on
three different levels of locality: the LBP labels for the
histogram contain information about the patterns on a
pixel-level, the labels are summed over a small region to
produce information on a regional level and the regional
histograms are concatenated to build a global description
of the face.
It should be noted that when using the histogram based
methods the regions do not need to be rectangular. Neither
do they requirebeing of the similar size or shape, nordo
they not necessarily have to cover the whole image. It is
also possible to have partially overlapping regions.
6. ILLUMINATION VARIATIONS
Illumination variation has enormously complex effects on
the image of an object. In the image ofa recognizable face,
shifting the direction of lightingdirect to shifts in the
location and figure ofshadows, transforms in highlights,
and exchange of difference gradients. Yet every- day
experienceshows that we are remarkably good at

www.ijsret.org

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 8, November 2014

1220

recognizing faces despite such variations in

lighting.Illumination can have a significant impact on the
form of surfaces, as the models ofshading, specularities
and shade change. For illustration, some pictures of a face
below differentlighting conditions. Differences in lighting
can often play a much greater rolein imagevariability of
human faces than differences between individual people.
Lighting designers inmovies can often set the mood of a
scene with carefully chosen lighting. To achieve a sinister
effect,for example, one can utilizelighting from below the
subject a sharp contrast to mostnaturalindoor or outdoor
scenes where the dominant light sources are above the
person.

to reduce the distance fromthe source to the reference

point cloud.
The ICP (Iterative Closest Point) algorithm is widely used
for geometricalignment of threedimensional models when
an initialestimate of the relative pose is well-known. A lot
of variants of ICPhavebeen anticipated, disturbing all
phases of the algorithm from the selectionand matching of
points to the minimization strategy.

Figure 4: Illumination image

Figure 6: ICP analysis

KRRR (KERNEL
REGRESSION)
7.

REDUCED

RANK

The basic idea is to use the KRRR model for describing

the relationship between the facialfeature vectors of
different facial views. In the KRRR model,the responses
are the featurevectors correspondingto another facial view
and the predictors are thefeature vectors from thegiven
facial view. By utilizingthe KRRR model, a feature vector
from any oneof the givenfacial views can be used to
estimate thefeature vectors of all other facial views.

Figure 5: KRRR analysis

8. ITERATIVE CLOSET POINT

Iterative Closest Point (ICP) is an algorithm employed
to minimize the difference between twoclosets of points.
ICP is frequentlyutilized to recreate 2D or 3D surfaces
from dissimilar scans, toconcentrate robots and
accomplishbest path planning, to co-register bone models,
etc.
In the algorithm aim is set asideset while the additional
one the basis is distorted to greatest match the orientation.
The algorithm iteratively adjusts thetransformation wanted

9. CONCLUSION:
Based on the assumption of a fully-controlled
environment for conscription, a face detection framework
is planned in which the widely-encountered single sample
problem for identification of faces with expressions is
targeted by augmenting the dataset with synthesized
images. Several expressions are simulated for each
enrolled person on an animatable model which is
specifically generated based on the 3D face scan of that
subject.Biometrics refers to the identification of humans
by their individuality or character. Biometrics is utilized in
computer science as a form of identification and access
control.Face detection in 3D images is not a common task.
Propose a novel framework for evaluating 3D faces, with
the exact goals of evaluating, corresponding, and
averaging their shapes.3D face scans affected due to
changes in facial expressions can also degrade face
recognition performance. Provide automatic detection of
17 landmarks is presented utilizing both 2D and 3D facial
data.Several expressions are simulated per persons and
recognition with real time datasets.Analyzed Receiver
Operating Characteristics (ROC) curves to visualize
verification performances with varying techniques.

REFERENCE
[1] L. D. Introna and H. Nissenbaum, Facial recognition
technology: A survey of policy and implementation
issues, in Report of the Center for Catastrophe
Preparedness and Response.New York, NY, USA: New
York Univ., 2009.

www.ijsret.org

International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 8, November 2014

[2] A. F. Abate, M. Nappi, D. Riccio, and G. Sabatino,

2D and 3D face recognition: A survey, Pattern Recognit.
Lett., vol. 28, no. 14, pp. 18851906, 2007.
[3] J. P. Phillips, P. Grother, R. J. Michaels, D. M.
Blackburn, E. Tabassi, and M. Bone, FRVT 2002
evaluation report, in Proc. IEEE Int. Workshop Anal.
Model. Faces Gestures, Oct. 2003, pp. 144.
[4] J. P. Phillips et al., FRVT 2006 and ICE 2006 largescale experimental results, IEEE Trans. Pattern Anal.
Mach. Intell., vol. 32, no. 5, pp. 831846, May 2009.
[5] W. Y. Zhao and R. Chelappa, SFS based view
synthesis for robust face recognition, in Proc. IEEE Int.
Conf. Autom. Face Gesture Recognit., Jan. 2000, pp. 285
292.
[6] X. Lu, R.-L. Hsu, A. K. Jain, B. Kamgar-Parsi, and B.
Kamgar-Parsi, Face recognition with 3D model-based
synthesis, in Proc. Int. Conf. Biometric Authentication,
2004, pp. 139146.
[7] Y. Hu, D. Jiang, S. Yan, L. Zhang, and H. Zhang,
Automatic 3D reconstruction for face recognition, in
Proc. IEEE Int. Conf. Autom. Face Gesture Recognit.,
May 2004, pp. 843848.
[8] M. W. Lee and S. Ranganath, Pose-invariant face
recognition using a 3D deformable model, J. Pattern
Recognit., vol. 36, no. 8, pp. 18351846, 2003.
[9] J. Huang, B. Heisele, and V. Blanz, Componentbased face recognition with 3D morphable models, in
Proc. Int. Conf. Audio Video-Based Biometric Person
Authentication, 2003, pp. 2734.
[10] U. Prabhu, J. Heo, and M. Savvides, Unconstrained
pose-invariant face recognition using 3D generic elastic
models, IEEE Trans. Pattern Anal. Mach. Intell., vol. 33,
no. 10, pp. 19521961, Oct. 2011.

www.ijsret.org

1221