Face recogntion using PCA algorithm

FACE RECOGNTION USING PCA ALGORITHM
Abstract
As society becoming more and more electronically connected, the capability to
automatically establish an identity of individuals using face as a biometric has become
important. Many applications such as identity verification, criminal face recognition, and
surveillance require robust and accurate face recognition technology. Face recognition has
become a very challenging problem in presence of clutter and variability of the background,
noise and occlusion, and finally speed requirements. This paper focuses on developing a face
recognition system using an extended PCA algorithm. The proposed algorithm uses the concept
of PCA and represents an improved version of PCA to deal with the problem of orientation and
lightening conditions present in the original PCA. The preprocessing phase of the proposed
algorithm emphasize the efficiency of the algorithm even when number of images per person or
the orientation is very different.
Keywords: Face recognition, Principal component Analysis

Table of Contents
Title
1.0 Introduction
2.0 Liturature and Survey and Background Knowledge
2.1 Face Recognition
2.2 Face Recognition using PCA algorithm
2.3 IPCA algorithm
2.4 Result and Discussion
3.0 Advantage and Limitations
3.1 Compared to other technologies
3.1.1Weaknesses
3.1.2Effectiveness
3.1.3Privacy issue
4.0 Applications
4.1 Access controll
4.2 Identification system
4.3 Surveillance
4.4 Pervasive Computing
5.0 Conclusion
References

1.0 INTRODUCTION
Face recognition has been a challenging and quite interesting problem in the field of
pattern recognition for a very long time. Beginning with Bledsoe's [2] and Kanade's [3] early
systems, a number of automated or semi automated face recognition strategies have modeled and
classified faces based on normalized distances and ratios among feature points. Recently this
general approach has been continued and improved by the recent work of Yuille et al [4].
Face recognition has received significant attention in the past decades due to its potential
applications in biometrics, information security, law enforcement, etc. By far, numerous methods
have been suggested to address this problem [3].
Digital images and video are becoming more and more important in the multimedia
information era. The human face is one of the most important objects in an image or video.
Detecting the location of human faces and then extracting the facial features in an image is an
important ability with a wide range of applications, such as human face recognition, surveillance
systems, human computer interfacing, video-conferencing, etc. In an automatic face recognition
system [1], the "rst step is to segment the face in an image or video irrespective of whether the
background is simple or cluttered. For model-based video coding [2], the synthesis performance
is quite dependent on the accuracy of the facial feature extraction process. In other words, a
reliable method for detecting the face regions and locating the facial features is indispensable to
such applications. This paper presents an efficient method for face detection and facial feature
extraction in a cluttered image.

2.0 Liturature and Survey and
Background Knowledge
2.1 FACE RECOGNTION
A facial recognition system is a computer application for automatically identifying or
verifying a person from a digital image or video frame from a video source . One of the ways to
do this is by comparing selected facial features from the image and a facial database .
It is typically used in security system and can be compared to other biometrics such as
fingerprint or eye iris recognition system[6].
Some facial recognition algorithms identify facial features by extracting landmarks, or
features, from an image of the subject's face. For example, an algorithm may analyze the relative
position, size----, and/or shape of the eyes, nose, cheekbones, and jaw [2] These features are then
used to search for other images with matching features [3] Other algorithms normalize a gallery
of face images and then compress the face data, only saving the data in the image that is useful
for face recognition. A probe image is then compared with the face data [4] One of the earliest
successful system [5] is based on template matching technique[6] applied to a set of salient facial
features, providing a sort of compressed face representation.
Recognition algorithms can be divided into two main approaches, geometric, which looks
at distinguishing features, or photometric, which is a statistical approach that distills an image
into values and compares the values with templates to eliminate variances.
Popular recognition algorithms include Principal Component Analysis using Eigen faced
,Linear Discriminate Analysis , Elastic Bunch Graph Matching using the Fisher face algorithm ,
the Hidden Markov Model, the Multilinear Subspace Learning using tensor representation , and
the Neuronal motivated dynamic link matching .

2.1.1 SOFTWARES BASED ON FACE RECOGNITION :
Notable software with face recognition ability include :
1] digiKam (KDE)
2] iPhoto (Apple)
3] OpenCV (Open Source)[10]
4] Photoshop Elements (Adobe Systems)
5] Picasa (Google)
6] Picture Motion Browser (Sony)
7] Windows Live Photo Gallery (Microsoft)
8] Mathematica
2.2 FACE RECOGNITION USING PCA ALGORITHM
Principal component analysis (PCA) turns out to be very effective [2]. Recently, a PCA
closely-related method, independent component analysis (ICA) [3], has also been applied to face
recognition. ICA can be viewed as a generalization of PCA since it concerns not only second-order
dependencies but also high-order dependencies between variables. The previous
researchers [4,5] however, all use the standard PCA as the baseline algorithm to evaluate ICA-based
face recognition systems. The initial success of eigenfaces popularized the idea of
matching images in compressed subspaces. Researchers began to search for other subspaces that
might improve performance. One alternative is Fisher's linear discriminant analysis (LDA, a.k.a.
"fisher faces") [6]. For any N-class classification problem, the goal of LDA is to find the N-1
basis vectors that maximize the interclass distances while minimizing the intraclass distances. At
one level, PCA and LDA are very different: LDA is a supervised learning technique that relies
on class labels, whereas PCA is an unsupervised technique. Nonetheless, in circumstances where

class labels are available either technique can be used, or LDA has been compared to PCA in
several studies [7]. Principal Component Analysis is a standard technique used to approximate
the original data with lower dimensional feature vectors [8]. The basic approach is to compute
the eigenvectors of the covariance matrix, and approximate the original data by a linear
combination of the leading eigenvectors. The mean square error (MSE) in reconstruction is equal
to the sum of remaining Eigen values. The feature vector here is the PCA projection coefficient.
PCA is appropriate when the samples are from one class or group (super class). In real
implementation, there are two ways to compute the eigenvalues and eigenvectors: SVD
decomposition and regular Eigen computation. For efficient way to compute or update the SVD,
please refer to [10]. In many cases, even though the matrix is full rank matrix, the large condition
number will create a numerical problem.
The distance measure used in the matching could be a simple Euclidean or a weighted
Euclidean distance. It has been suggested that the weighted Euclidean will give better
classification than the simple Euclidean distance [10]. Moreover this technique can also be
applied for the purpose of Facial Expression Analysis. Most approaches to automatic facial
expression analysis attempt to recognize a small set of prototypic emotional facial expressions,
i.e., fear, sadness, disgust, anger, surprise, and happiness (e.g., [9, 10]. This practice follows
from the work of Darwin [10], and more recently Ekman [8], who suggested that basic emotions
have corresponding prototypic expression .
2.3 IMPROVED PRINCIPLE COMPONENT ANALYSIS (IPCA)
Principal Component Analysis (PCA) is a way of identifying patterns in data, and
expressing the data in such a way as to highlight their similarities and differences. In PCA the
variance of each image form the mean image is determined. This variance is a measure of
variability in the face space. To apply PCA for face recognition, Eigenfaces are calculated and
the weight of these Eigenfaces is used to find the contribution of training images to the input
image. In the proposed work, certain changes to the original PCA algorithm are made. The
preprocessing of the training images has been done to remove the background, lightening
conditions and the orientation factors. Also, some normalization steps have been included to

remove the calculation induced errors. Our proposed algorithm is an improvement of PCA
algorithm. The original PCA algorithm didn't work well when the orientation of the images was
very large i.e. around 90 degrees. But our proposed algorithm worked quite well even in those
cases in which PCA failed.
2.3.1 The main steps of Improved PCA algorithm are:
1] The faces in the training set are preprocessed by taking the co-ordinates of both the eyes
and mouth and then applying cropping and aligning on these distances.
2] From the preprocessed training set, compute the Eigenfaces and then obtain the best
Eigenfaces corresponding to highest Eigen values.
3] Now project these Eigenfaces into the face database to find their contribution.
4] Take an input image which has to be identified and apply the same preprocessing steps.
5] Find the weight pattern of the Eigenfaces by projecting the input image into Eigen faces.
6] Now reconstruct the input image from the weighted Eigenfaces.
7] Determine if it is a face at all and if so, either known or unknown.
2.4 RESULTS AND DISCUSSION
The experiments were conducted in MATLAB 7.0 taking following parameters: A total
of 110 images were taken in the training set. There were 11 images per person and a total of 10
persons. These 11 images were in the basic 11 directions including the 90 degree orientation.
These images were taken from FERET database which is a standard database meant for testing
various face recognition algorithms. These images were then preprocessed and stored at a
separate place. The input image was also preprocessed to remove the noise from it. The

resolution of the original images was 256*384. After the preprocessing the resolution became
241*291 since only face part of the image was taken.
2.4.1 The two test cases were classified as follows:
1] As the first test case, we took the image of the person who was in the database but the input
image was not included in the database. It had a different orientation which we have taken as 90
degree to test the robustness and efficiency of the algorithm.
2] As the second case, we took the image of the person which was exactly present in the
database.
The Figure 1(b) shows the preprocessing image of input image as shown in Figure 1(a).
Fig. 1 (a) Input Image Fig 1. (b) Preprocessed Image.
2.4.2 Experiment 1: A different image taken as input
As explained above, we took the image of the person who was in the database but the
input image was not included in the database. It had a different orientation, which we have taken
as 90 degree, to test the robustness and efficiency of the algorithm.

Now our algorithm reconstructs the input image by considering the weights if the
Eigenfaces and the contribution of each face. The reconstructed image is shown to the right part
of Figure 2.
Fig. 2 (a) Left part shows input image fig.2. (b) Right part shows Reconstructed Image.
It is evident from Figure 2 , the reconstructed image resembles the preprocessed input
image hence improving the recognition efficiency . The plot of the weight of the input image
against face space and the Euclidean distance of the input image from all of the face space
images is shown in Figure 3. It is clear that the Euclidean distance of image "103.jpg" is the
lowest and below the threshold value. Also the Euclidean distances of the input image form the
face class (all the images of a single person) are comparable.
The face identified by the proposed algorithm is shown in Figure 3a
Fig. 3a. Face identified from database.

2.4.3. Experiment 2:
Input image present exactly in the database
As the second case, we took the image of the person which was exactly present in the
database as shown in Figure 4 (a). The image reconstructed by the algorithm is shown in Figure
4(b).
Fig. 4. (a) Left part shows input image (b) Right part shows Reconstructed Image.
As the input image is exactly same as in the database, so the reconstructed image is very
fine. Now in Figure 6 shown below, we have plotted the Euclidean distance of the input image
from all of the images in the face space. As the image was present exactly, so the Euclidean
image of the same image i.e. "12.jpg" is zero. This is also true theoretically.
The face identified by the proposed algorithm is shown in Figure 5
Fig. 5a. Face identified from database

3.0 ADVANTAGE AND LIMITATIONS
3.1 Compared to other technologies:
Among the different biometric techniques, facial recognition may not be the most reliable
and efficient .However, one key advantage is that it does not require the cooperation of the test
subject to work. Properly designed systems installed in airports, multiplexes, and other public
places can identify individuals among the crowd, without passers-by even being aware of the
system. Other biometrics like fingerprints, iris scans, and speech recognition cannot perform this
kind of mass identification. However, questions have been raised on the effectiveness of facial
recognition software in cases of railway and airport security.
3.1.1 Weaknesses:
Face recognition is not perfect and struggles to perform under certain conditions. Ralph
Gross, a researcher at the Carnegie Mellon Robotics Institute, describes one obstacle related to
the viewing angle of the face: "Face recognition has been getting pretty good at full frontal faces
and 20 degrees off, but as soon as you go towards profile, there've been problems.
Other conditions where face recognition does not work well include poor lighting,
sunglasses, long hair, or other objects partially covering the subject’s face, and low resolution
images[7].
Another serious disadvantage is that many systems are less effective if facial expressions
vary. Even a big smile can render the system less effective. For instance: Canada now allows
only neutral facial expressions in passport photos[9]
There is also inconstancy in the datasets used by researchers. Researchers may use
anywhere from several subjects to scores of subjects, and a few hundred images to thousands of
images. It is important for researchers to make available the datasets they used to each other, or
have at least a standard dataset[10]. On 18 January 2013 Japanese researchers created a privacy

visor that uses nearly infrared light to make the face underneath it unrecognizable to facial
recognition software[11].
3.1.2 Effectiveness:
Critics of the technology complain the London Borough of Newham scheme has, as of
2004, never recognized a single criminal, despite several criminals in the system's database
living in the Borough and the system having been running for several years. "Not once, as far as
the police know, has New ham’s automatic facial recognition system spotted a live target.This
information seems to conflict with claims that the system was credited with a 34% reduction in
crime (hence why it was rolled out to Birmingham also).However it can be explained by the
notion that when the public is regularly told that they are under constant video surveillance with
advanced face recognition technology, this fear alone can reduce the crime rate, whether the face
recognition system technically works or does not. This has been the basis for several other face
recognition based security systems, where the technology itself does not work particularly well
but the user's perception of the technology does.
An experiment in 2002 by the local police department in Tampa, Florida, had similarly
disappointing results. A system at Boston's Logan Airport was shut down in 2003 after failing to
make any matches during a two-year test period
3.1.3 Privacy issues:
Civil rights right organizations and privacy campaigners such as the EFF and
the ACLU express concern that privacy is being compromised by the use of surveillance
technologies. Some fear that it could lead to a “total surveillance society,” with the government
and other authorities having the ability to know the whereabouts and activities of all citizens
around the clock. This knowledge has been, is being, and could continue to be deployed to
prevent the lawful exercise of rights of citizens to criticize those in office, specific government
policies or corporate practices. Many centralized power structures with such surveillance

capabilities have abused their privileged access to maintain control of the political and economic
apparatus, and to curtail populist reforms [10].
Facial recognition can be used not just to identify an individual, but also to unearth other
personal data associated with an individual – such as other photos featuring the individual, blog
posts, social networking profiles, Internet behavior, travel patterns, etc. – all through facial
features alone [10].Moreover, individuals have limited ability to avoid or thwart facial
recognition tracking unless they hide their faces. This fundamentally changes the dynamic of
day-to-day privacy by enabling any marketer, government agency, or random stranger to secretly
collect the identities and associated personal information of any individual captured by the facial
recognition system [4].
Social media web sites such as Facebook have very large numbers of photographs of
people, annotated with names. This represents a database which could be potentially used (or
abused) by governments for facial recognition purposes.
In July 2012, a hearing was held before the Subcommittee on Privacy, Technology and
the Law of the Committee on the Judiciary, United States Senate, to address issues surrounding
what facial recognition technology means for privacy and civil liberties.

4.0 APPLICATION
Many applications for face recognition have been envisaged, and some of them have been
hinted at above. Commercial applications have so far only scratched the surface of the potential.
Installations so far are limited in their ability to handle pose, age and lighting variations, but as
technologies to handle these effects are developed, huge opportunities for deployment exist in
many domains.
4.1 Access Control:
Face verification, matching a face against a single enrolled exemplar, is well within the
capabilities of current Personal Computer hardware. Since PC cameras have become widespread,
their use for face-based PC logon has become feasible, though take-up seems to be very limited.
Increased ease-of-use over password protection is hard to argue with today’s somewhat
unreliable and unpredictable systems, and for few domains is there motivation to progress
beyond the combinations of password and physical security that protect most enterprise
computers. As biometric systems tend to be third party, software add-ons the systems do not yet
have full access to the greater hardware
security guarantees afforded by boot-time and hard disk passwords. Visionics’ face-based screen
lock is one example, bundled with PC cameras. Naturally such PC-based verification systems
can be extended to control authorization for single-sign-on to multiple networked services, for
access to encrypted documents and transaction authorization, though again uptake of the
technology has been slow.
Face verification is being used in kiosk applications, notably inMr. Payroll’s (now
Innoventry) cheque-cashing kiosk with no human supervision. Innoventry claims to have one
million enrolled customers. Automated TellerMachines, already often equipped with a camera,
have also been an obvious candidate for face recognition systems (e.g. Viisage’s FacePIN), but
development seems not to have got beyond pilot schemes. Banks have been very conservative in

deploying biometrics as they risk losing far more through customers disaffected by being falsely
rejected than they might gain in fraud prevention. Customers themselves are reluctant to incur
burdensome additional security measures whentheir personal liability is already limited by law.
For better acceptance, robust passive acquisition systems with very low false rejection
probabilities are necessary.
Physical access control is another domain where face recognition is attractive (e.g. Cognate’s
FaceVACS, Miros’ TrueFace) and here it can even be used in combination with other biometrics.
BioId [3] is a system which combines face recognition with speaker identification and lip
motion.
4.2 Identification Systems:
Two US States (Massachusetts and Connecticut [3]) are testing face recognition for the
policing of Welfare benefits. This is an identification task, where any new applicant being
enrolled must be compared against the entire database of previously enrolled claimants, to ensure
that they are not claiming under more than one identity. Unfortunately face recognition is not
currently able to reliably identify one person among the millions enrolled in a single state’s
database, so demographics (zip code, age, name etc. ) are used to narrow the search (thus
limiting its effectiveness), and human intervention is required to review the false alarms that
such a system will produce. Here a more accurate system such as fingerprint or iris-based person
recognition is more technologically appropriate, but face recognition is chosen because it is more
acceptable and less intrusive. In Connecticut, face recognition is the secondary biometric added
to an existing fingerprint identification system. Several US States, including Illinois, have also
instituted face recognition for ensuring that people do not obtain multiple driving licenses.
4.3 Surveillance:
The application domain where most interest in face recognition is being shown is
probably surveillance. Video is the medium of choice for surveillance because of the richness

and type of information that it contains and naturally, for applications that require identification,
face recognition is the best biometric for video data. Though gait or lip motion recognition have
some potential. Face recognition can be applied without the subject’s active participation, and
indeed without the subject’s knowledge. Automated face recognition can be applied ‘live’ to
search for a watch-list of ‘interesting’ people, or after the fact using surveillance footage of a
crime to search through a database of suspects. The deployment of face-recognition surveillance
systems has already begun
, though the technology is not accurate enough yet [4]. The US government is investing in
improving this technology [10] and while useful levels of recognition accuracy may take some
time to achieve, technologies such as multiple steerable zoom cameras, non-visible wavelengths
and advanced signal processing are likely to bring about super-human perception in the data
gathering side of surveillance systems.
4.4 Pervasive Computing:
Another domain where face recognition is expected to become very important, although
it is not yet commercially feasible, is in the area of pervasive or ubiquitous computing. Many
people are envisaging the pervasive deployment of information devices. Computing devices,
many already equipped with sensors, are already found throughout our cars and in many
appliances in our homes, though they will become ever more widespread. All of these devices
are just now beginning to be networked together. We can envisage a future where many
everyday objects have some computational power,
Allowing them to adapt their behavior —to time, user, user control and a host of other factors.
The communications infrastructures permitting such devices to communicate to one another are
being defined and developed (e.g. Bluetooth, IEEE 802.11). So while it is easy to see that the
devices will be able to have a well-understood picture of the virtual world with information
being shared among many devices, it is less clear what kind of information these devices will
have about the real physical world. active commands on the part of the user. Some simple
devices can sense the environment, but it will be increasingly important for such pervasive,
networked computing devices to know about the physical world and the people within their
region of interest. Only by making the pervasive infrastructure ‘human aware’ can we really reap

the benefits of productivity, control and ease-of-use that pervasive computing promises. One of
the most important parts of human awareness is knowing the identity of the users close to a
device, and while there are other biometrics that can contribute to such knowledge, face
recognition is the most appropriate because of its passive nature. There are many examples of
pervasive face recognition tasks: Some devices such as Personal Digital Assistants (PDAs) may
already contain cameras for other purposes, and in good illumination conditions will be able to
identify their users. A domestic message centre may have user personalization that depends on
identification driven by a built-in camera. Some pervasive computing environments may need to
know about users when not directly interacting with a device, and may be made ‘human aware’
by a network of cameras able to track the people in the space and identify each person, as well as
have some understanding of the person’s activities. Thus a video conference room could steer the
camera and generate a labeled transcript of the conference; an automatic lobby might inform
workers of specific visitors; and mobile workers could be located and kept in touch by a system
that could identify them and redirect phone calls.

5.0 CONCLUSION
This paper presents an algorithm to recognize faces present in the face database. The
proposed algorithm uses the concept of PCA and represents an improved version of PCA to deal
with the problem of orientation and lightening conditions present in the original PCA. In the case
when test image was in the database, the person was identified correctly. Now even when we
took as input, an image with different orientation of a person present in database, the algorithm
successfully identified the person. This shows that pre-processing greatly enhances the efficiency
of the algorithm even when we have less number of images per person or the orientation is
greatly different. This work is being extended to deal with a range of aspects (other than full
frontal views) by defining a small number of classes for each known person corresponding to
characteristic views.

REFERENCES
[1] K.-M. Lam, H. Yan,” An analytic-to-holistic approach for face recognition based on a single
frontal view”, IEEE Trans. Pattern Anal. Mach. Intell. 20 (1998) 673-686.
[2] L. Zhang,” Automatic adaptation of a face model using action units for semantic coding of
videophone sequences”, IEEE Trans. Circuits Systems Video Technol. 8 (6) (1998) 781-795.
[3] W. Zhao, R. Chellappa, A. Rosenfeld, and P. Phillips (2002) ,” Face recognition: A literature
survey.” Technical Report CAR-TR-948, UMD CS-TR-4167R, August, 2002
[4] W. W. Bledsoe (1966),” The model method in facial recognition Panoramic Research “Inc.,
Palo Alto, CA, and Rep. PRI: 15, Aug. 1966
[5] T. Kanade (1973),” Picture processing system by computer complex and recognition of
human face” , Dept. of Information Science, Kyoto University, Nov. 1973
[6] A.L. Yuille, D.S. Cohen, and P.W. Hallinan (19889) ,“Feature extraction from faces using
deformable templates”, proc. CVPR, San Diego, CA, June 1989.
[7] Bonsor, K., “How Facial System Works ” Retrieved 2008-06-02.
[9] Harley Geiger (2011-12-06). ,"Facial Recognition and Privacy". , Center for Democracy &
Technology. Retrieved 2012-01-10.
[10] What Facial Recognition Technology Means for Privacy and Civil Liberties: Hearing before
the Subcommittee on Privacy, Technology and the Law of the Committee on the Judiciary,
United States Senate, One Hundred Twelfth Congress, Second Session, July 18, 2012

Face recogntion using PCA algorithm

More Related Content

What's hot (20)

Viewers also liked (7)

Similar to Face recogntion using PCA algorithm (20)

More from Ashwini Awatare (9)

Recently uploaded (20)

Face recogntion using PCA algorithm