Image Procesing

							           ACUMEN-09

            Facial Recognition using the theory of biometrics

IMAGE PROCESSING using THEORY OF BIO-METRICS
Abstract:
                                              face and creates a faceprint, a unique
            While humans have had the         numerical code for that face. Once
innate ability to recognize and               the system has stored a face print, it can
distinguish different faces for millions of   compare it to the thousands or millions
years, computers are just now catching        of face prints stored in a database.
up.                                           Potential applications even include ATM
                                              and check-cashing security, Security
          In this paper, we'll learn how      Law      Enforcement       &     Security
computers are turning your face into          Surveillance and voter database for
computer code so it can be compared to        duplicates. This biometrics technology
thousands, if not millions, of other faces.   could also be used to secure your
We'll also look at how facial                 computer files. By mounting a Webcam
recognition software is being                 to your computer and installing the facial
used in elections, criminal investigations    recognition software, your face can
and to secure your personal computer.         become the password you use to get
Facial recognition software falls into a      into your computer. By implementing this
larger group of technologies known as         technology and the normal password
biometrics.                                   security you are getting double security
                                              to your valuable data.
         Biometrics uses biological
information to verify identity. The basic
idea behind biometrics is that our bodies
contain unique properties that can be
used to distinguish us from others.
Facial recognition methods may vary,
but they generally involve a series of
steps that serve to capture, analyze
and compare your face to a database
of stored images. A Software company
called Visionics developed Facial
Recognition software called Face it.


The heart of this facial recognition
system is the Local Feature
Analysis (LFA) algorithm. This is the
mathematical technique the system uses
to encode faces. The system maps the
Introduction                                  faces. While humans have had the
People have an amazing ability to             innate ability to recognize and
recognize and remember thousands of           distinguish different faces for millions of
years, computers are just now catching        faces for millions of years, computers
up. In this paper, you'll learn how           are just now catching up. Visionics, a
computers are turning your face into          company based in New Jersey, is one of
computer code so it can be compared to        many developers of facial recognition
thousands, if not millions, of other faces.   technology. The twist to its particular
We'll also look at how facial                 software, FaceIt, is that it can pick
recognition software is being                 someone's face out of a crowd, extract
used in elections, criminal investigations    that face from the rest of the scene and
and to secure your personal computer.         compare it to a database full of stored
                                              images. In order for this software to
The Face                                      work, it has to know what a basic face
Your face is an important part of who         looks like.
you are and how people identify you.
Imagine how hard it would be to
recognize an individual if all faces
looked the same. Except in the case of
identical twins, the face is arguably a
person's     most     unique    physical
characteristic.

While humans have had the innate
ability to recognize and distinguish
differ
ent
Facial recognition software can be used to find criminals in
a crowd, turning a mass of people into a big line up.

Facial recognition software is based on             Width of nose
the ability to first recognize a face, which        Depth of eye sockets
is a technological feat in itself, and then         Cheekbones
measure the various features of each                Jaw line
face. If you look in the mirror, you can            Chin
see that your face has certain                 These nodal points are measured to
distinguishable landmarks. These are           create a numerical code, a string of
the peaks and valleys that make up the         numbers that represents the face in a
different facial features. Visionics           database. This code is called a
defines these landmarks as nodal               faceprint. Only 14 to 22 nodal points are
points. There are about 80 nodal points        needed for the FaceIt software to
on a human face. Here are a few of the         complete the recognition process. In the
nodal points that are measured by the          next section,we'll look at how the system
software:                                      goes about detecting, capturing and
                                               storing faces.
●Distance between eyes




The Software

Facial recognition software falls into a            Fingerprint scan
larger group of technologies known as               Retina scan
biometrics.                                         Voice identification
                                               Facial recognition methods may vary,
Biometrics uses biological information to      but they generally involve a series of
verify identity. The basic idea behind         steps that serve to capture, analyze and
biometrics is that our bodies contain          compare your face to a database of
unique properties that can be used to          stored images. Here is the basic process
distinguish us from others. Besides            that is used by the FaceIt system to
facial        recognition,      biometric
                                               capture and compare images:
authentication methods also include:




To identify someone, facial recognition software compares
newly captured images to databases of stored images.
Detection - When the system is                      2. Normalization -The image
attached to a video surveillance system,     of the head is scaled and rotated so that
the recognition software searches the        it can be registered and mapped into an
field of view of a video camera for faces.   appropriate       size      and       pose.
If there is a face in the view, it is        Normalization is performed regardless of
detected within a fraction of a second. A    the head's location and distance from
multi-scale algorithm is used to             the camera. Light does not impact the
search for faces in low resolution. (An      normalization process.
algorithm is a program that provides a              3. Representation - The
set of instructions to accomplish a          system translates the facial data into a
specific task). The system switches to a     unique code. This coding process allows
high-resolution search only after a head-    for easier comparison of the newly
like shape is detected.                      acquired facial data to stored facial data.
                                                    4. Matching - The newly
      1. Alignment - Once a face is          acquired facial data is compared to the
detected, the system determines the          stored data and (ideally) linked to at
head's position, size and pose. A face       least one stored facial representation.
needs to be turned at least 35               The heart of the FaceIt facial recognition
degrees toward the camera for the            system is the Local Feature
System to register it.                       Analysis (LFA) algorithm. This is the
mathematical technique the system uses       from memory or 15 million per minute
to encode faces. The system                  from hard disk. As comparisons are
maps the face and creates a face             made, the system assigns a value to the
print, a unique numerical code for that      comparison using a scale of one to 10. If
face. Once the system has stored a face      a score is above a predetermined
print, it can compare it to the thousands    threshold, a match is declared. The
or millions of face prints stored in a       operator then views the two photos that
database. Each face print is stored as       have been declared a match to be
an 84-byte file.                             certain that the computer is
The system can match multiple face
prints at a rate of 60 million per minute
accurate. Facial recognition, like other
form                                    of
biometrics, is considered a technology       at catching these people. Using the
that will have many uses in the near         facial recognition technology, officials
future. In the next section, we will look    can search through facial images in the
how it is being used right now. The          voter database for duplicates at the time
primary users of facial recognition          of registration. New images are
software like Face It have been law          compared to the records already on file
enforcement agencies, which use the          to catch those who attempt to register
system to capture random faces in            under aliases. Potential applications
crowds. These faces are compared to          even include ATM and check-cashing
Database Image to a database of              security. The software is able to quickly
criminal mug shots. In addition to law       verify a customer's face. After the user
enforcement and security surveillance,       consents, the ATM or check-cashing
facial recognition software has several      kiosk captures a digital photo of the
other uses, including:                       customer. The facial recognition
     Eliminating voter fraud                software then generates a face print of
     Check-cashing              identity    the photograph to protect customers
         verification                        against identity theft and fraudulent
     Computer security                      transactions. By using facial recognition
One of the most innovative uses of facial    software, there's no need for a picture
recognition is being employed by the         ID, bank card or personal identification
Mexican government, which is using the       number (PIN) to verify a customer's
technology to weed out duplicate             identity.
voter registrations. To sway an              Many people who don't use
election, people will register several       banks use check-cashing
times under different names so they can      machines.
vote more than once. Conventional
methods have not been very successful
Facial recognition could                  and installing the facial recognition
eliminate possible criminal               software, your face can become the
activity.                                 password you use to get into your
This biometric technology could also be   computer. IBM has incorporated the
used to secure your computer files. By    technology into a screensaver for it’s A,
mounting a Webcam to your computer        T and X series ThinkPad laptops.


Webcam and Facial Recognition Software installed
Facial recognition software can be used to lock your computer.




Conclusion
With the following advantages and also       While facial recognition can be used to
some of the drawbacks, we conclude          protect your private information, it can
our paper on Facial Recognition using       just as easily be used to invade your
Biometrics. Potential applications are as   privacy by taking you picture when you
follows:                                    are entirely unaware of the camera. As
      Eliminating voter fraud              with many developing technologies, the
      Security law enforcement             incredible potential of facial recognition
       and Security surveillance            comes with drawbacks. But if we add
      ATM and Check-cashing                both the facial recognition and the
       identity verification                normal password security we can have
      Computer security                    an added Double Security which is
                                            more reliable than one shield security,
                                            Just same as the quote “Two heads
                                            are better than one”.




References:

1. www.google.com\image processing
2. www.howstuffworks.com\elecronics stuff
3. IEEE Journals on Image Processing