Seminar Report

SEMINAR REPORTFace recognition system using eigenfaces

VIHANG PATTIWAR ELECTRONICS (BE)

Face recognition system using eigenfaces

BANSILAL RAMNATH AGARWAL CHARITABLE TRUST’S

VISHWAKARMA INSTITUTE OF TECHNOLOGY(An Autonomous Institute affiliated to University of Pune)

PUNE - 411037

CERTIFICATE

This is to certify that the Seminar Report titled

“Face recognition system using eigenfaces”

___________________________________________________________________

has been submitted in the academic year 2010 / 2011 by

Vihang pattiwar(BE F 05)

____________________________________________________________________

In partial fulfillment of the Bachelor’s Degree in the ELECTRONICS Engineering

as prescribed by the University of Pune.

Date: Guide Head of the Department

Vit Electronics Page no 2


ACKNOWLEDGMENT

I acknowledge a deep sense of gratitude towards PROF. S. B. BHISE, seminar guide and assistant

professor of electronics department, V.I.T. Pune , for providing me valuable and useful information with

guidance and co-operation in writing this report. I would like to thank all other professors and my

friends for giving help and encouragement in all the problems faced during preparation of this paper. I

would like to thank all those who are indirectly involved in preparation of this paper- my colleagues,

friends, parents.

Vihang pattiwar

(BE F)



INDEX: PAGE NO.

ABSTRACT 5

1. Introduction

1.1 Introduction 6

1.2 overview 7

1.3 Importance of face recognition 8

2. Face recognition system

2.1 Approaches for face recognition. 9

2.2 Block diagram of face recognition system 10

2.3 Comparisons’ of eigenface approach to feature based approach. 13

2.4 Difficulties 14

3. Face recognition using eigenface

3.1 Introduction 15

3.2 Overview over the algorithm 18

3.3 Eigenvectors and eigenvalues 19

3.4 Calculation of eigenfaces with PCA. 20

3.5 Merits and demerits 26

3.6 Application 27

4. SOME IMAGES OF FACE RECOGNITION 28

REFERENCES 30



ABSTRACT

The rapid development in information technology has made use of computers mandatory

in each and every field.

Digital image processing is a rapidly evolving field with growing applications in science

and engineering . Image processing holds the probability of developing the ultimate machine

that could perform the visual function of all living beings.

In this project an approach is made to detect and identify a human face and describe

the algorithm for software implementation of face recognition system using eigenface. In

eigenface method, training set is prepared first and then the person is recognized by comparing

characteristics of the face to those of known individuals.

Above technique is operated on BMP and JPEG image format. The software has been

developed in MATLAB.



CHAPTER 1:

1.1 INTRODUCTION:

Face is our primary focous of interaction with society, face communicates identify,

Emotion, race and age. It is also quite useful for judging gender, size and perhaps even character

Of the person.

A] IMAGE:

The term image is refers to two dimensional light intensity function denoted by

f(x,y)Wher value or amplitude of “f “ at spatial co-ordinates (x,y) gives the intensity of image at

that point . as light is a form of energy f(x,y) be non zero and finite, that is

0< f(x,y) < ∞

B] IMAGE PROCESSING:

The image processing is the science of manipulating a picture. It covers a broad range

Of techniques that are present in numerous applications. This technique can enhance or distort

an image, create a new image that has been degraded during or after the image acquition.

Certainly the most powerful image processing system we see and use every day is the one

compose of the human eye, brain and face. Image processing has grown enormously and forced

It into number of fields where cameras amd image are involved. Digital television, computer

axial tomography (in medicine) are some of the few other areas where digital image processing

(DIP) touches the life of common man.



1.2 OVERVIEW OF FACE RECOGNITION:

Developing a computational model for face recognition is quite difficult, because faces

are too complex, multidimensional, and meaningful visual stimuli. They are natural class of

objects, and stand in stark contrast to sine wave greetings, the “block words”, and other artificial

stimuli used in human and computer vision research. Thus unlike most early visual, for which we

may construct detailed models of retinal or striate activity, face recognition is a very high level

task for which computational approaches can currently only suggest broad constraints on the

corresponding neural activities .

Taking the advantage of some of this structure by proposing a scheme for recognition

Which based on an information theory,seeking to encode the most relevant information in a

group of faces which will best distinguish them from one another. This transforms face image

into a small set of characteristic feature images called “eiganfaces” which are the principal

componantes of the initial training set of face image. Recognition is performed by projecting a

new image into subsequence spanned by the eiganfaces and then classifying the face by

comparing its portion of known individuals.

Automatically learning and later recognizing new faces is practical within this framework

Recognition under reasonably varying conditions is achieved by training or a limited number of

characteristics views. This has advantage over other face recognition schemes in its speed and

simplicity, learning capacity and relative insensitivity to small or gradual change in the face

image.



1.3 IMPORTANCE OF FACE RECOGNITION:

In today’s network world, the need to maintain the security of information or physical

property is becoming both inceasingly difficult. From time to time we hear about the crimes

Like credit card frauds, network intrusions, or security branches.

The criminals are taking advantage of a fundamental flaw in the conventional access

control systems.

The system do not grant access by “who we are”, but by “what we have”, such as ID

cards , keys , passwords etc. none of these means actually define us. Rather , they merely

Are means to authenticate us. It goes without that if someone steals duplicates or acquires these

Identity means, he or she will be able to access our data or our personal property any time and

anywhere.

Recently, technology became available to allow verification of “true” individual identity.

This technology is based on a field called “biometrics”. Biometric access control methods are

automated methods of verifying or recognizing the identity of a person o the basis of some

physiological characteristics, such as fingerprints or facial features , or some aspects of the

person’s beheviour , like handwriting style or keystroke patterns.since biometric systems identify

a person by his/her biological characteristics, they are difficult to forge.

Among the various biometric ID methods, the physiological methods (fingerprint,

faceDNA) are more stable than methods in behavioral category (keystroke, voiceprint). The

reason is that physiological features are often non-alterable except by severe injury. The

behavioral pattern, on the other hand, may fluctuates due to stress, fatigue or illness. However,

Behavioral IDs have the advantage of being non-intrusive.people are more comfortable signing

Their names or speaking to a microphone than placing their eyes before a scanner or giving a

drop for DNA sequencing. Face recognition is one of the few biometric methods that possess

The merits of both high accuracy and low intrusiveness.



CHAPTER 2:FACE RECOGNITION SYSTEM:

2.1 APPROACHES FOR FACE RECOGNITION:

The major approaches used for face recognition are-

1) Featured based approach

2) Eiganface based approach

1] Feature based approach:

It is based on extraction of properties of individual organs located on a face such

as eyes, nose and mouth as well as their relationships with each other. Effective features

That can be used for face recognition can be classified as follows:

First order features values

Discrete feature such as eyes, eyebrows, mouth, chin and nose which have been

found to be important in face identification and are specified without reference to other

facial features, are called first-order features.

Second order features values

Another configurable set of features that characterize the spatial relationship

between the position of first order features and information about the shape of face

Are called second order features. This approach is simple to implement and heuristics

Can be applied easily but it is not a robust method for face recognition. This method

Is heavily dependent on the external environmental features.



2]EIGENFACE BASED APPROACH:

Eigenface based approach is derived from information theory that extract the relevant

information, encode it as efficiently as possible and compare one encoding with database of

models encoded similarly. A simple approach to extract the information contained in the image

of face is to capture the variation of images, dependant of judgment of features and use this

information to encode & compare individual face image. In eigen face based approaches,

principle component analysis is used to encode the information.

2.2 BLOCK DIAGRAM OF FACE RECOGNATION SYSTEM:

The face recognition system passes through three main phases during face

recognition process. Three major functional units are involved in three phases and they are

depicted in figure.

Face library formation phase :

In this phase, the acquisition and the pre preprocessing of face images that are

going to be added to face library are performed. Face images are stored in a face library

in the system. The face database is called as “face library” because at the moment it does

not have the properties of relational database. Every action such as training set or eigen

face formation is performed on face library. In order to start the face recognition process,

this initially empty face library has to be filled with face images. The face recognition

system operates on 256*256 BMP formatted images files. In order to perform image size

conversions and enhancement on face images their exists the “pre-processing” module.

This module automatically converts every face image to 256*256 and based on user

request, it can modify the dynamic range of face images (histogram equalization) in order

to improve face recognition performance. Also implement a “background removal”



algorithm in the preprocessing module. Each face is represented by one entry in the face

library. The entry corresponds to the face image itself for the sake of speed, no data

compression is performed on face image that is stored in face library.

Training phase:

After adding the face images to face library, the system is ready to perform

training set and eigen face formation. Those face images that are going to be in the

training set are chosen from the entire face library. Because that the face library enties are

normalized, no further preprocessing is necessary at this stage. After choosing the

training set einen faces are formed and stored for later use. Eigen faces are calculated

from training set keeping only the M images that corresponds to highest eigen values.

This M eigen faces define the M dimensional “face space”. As new faces are

experienced, the eigen faces can be updated or calculated. the corresponding distribution

in the M dimensional weight space is calculated for each face library member by

projecting its face image on to the “face space” spanned by the eigen space. Now the

corresponding weight vector of each face library member has been updated which were

initially empty. The system is now ready for recognition process. Once a training set has

been chosen it is not possible to add new member to the library with the conventional

method that is presented in phase one. Because, the system does not know already exist

in face library or not a library search must be performed.

RECOGNITION AND LEARNING PHASE:

After choosing a training set and constructing the weight vectors of face library

members , now the system is ready to perform the recognition process.user initiates the

recognition process by choosing a face image. Based on the user request and the acquired

image size, pre-processing steps are applied to normalize

This acquired image to face library specifications. Once the image is

normalized.,its weight vector is constructed with the help of the eigenface that are were

already stored during the training phase.after obtaining the weight vector,it is compared

with the weight vector of every face member within the a user defined “threshold”.if



there exist at least one face library member that is similar to the acquired image within

that threshold then, the face image is classified as “known”.otherwise, a miss has

occurred and the face is classified an “unknown”



Block Diagram of Face recognition system



2.3 COMPARISION OF THE EIGENFACE TO FEATURE APPROACH:

These two different approaches are compared based on the following aspects of face

recognition.

Speed and simplicity

Feature based face recognition involves complex computations such as deformable

templates and active contour models. The evaluations of these parameters are very time

consuming even on today’s computers. Eigenfaces approach is superior in its speed and

reasonably simple impletation, it is keen that the evaluation of a feature vector involves

merely additions and multiplications. On a machine that is capable of executing an

addition and multiplication in one clock cycle, this feature evalution and comparision can

be done in real time.during the experiments, although no special hardware (and special

Software optimizations for matrix operations) was used, speed of the proposed face

recognition system was near real time.

Learning capability

Feature based face recognition systems are generally trained to optimize their parameters

In supervised manner. That is the system designer presents known individuals to the

system and checks systems response. In the Eigenface approach,training is done in an

unsupervised manner. User selects training set that represents the rest of the face images .

Eigenface are obtained from the training set members and feature vectors are formed.

Face background

Backgrounds of the face images are extremely important in the eigenface approach.

Eigenface and feature vectors are evaluated by image multiplication and addition. As a

result of this, entire information contained in the face image is used. If this information

changes due to face background, recognition performance can significantly decrease. In

order to avoid this a “back-ground removal” algorithm can be less sensitive to face

background in case. They are generally locate face contours In order to extract facial

features.



2.4DIFFICULTIES:

Due to the dynamic nature of face images, a face recognition system encounters

various problems during the recognition process. It is possible to classify a face

recognition system as either “roboust” or “weak” based on its recognition performances

under these circumstances. The objectives of a roboust face recognition is given below.

SHIFT INVARIENCE

The same can be to the system at different perspectives and orientations. For

instances , face images of the same person could be taken from frontal and profile

views. Besides head orientation may change due to translation and rotations.

SCALE INVARIENCE

The same face can be presented to the system at different scales. This may happen

due to focal distance between the face and the camera. as this distance gets

closer , the face images get closer.

ILLUMINATION INVARIENCE

Face images of the persons can be taken under different illumination conditions

such as, the position and the strength of the light source.

EMOTIONAL EXPRESSION AND DETAIL INVARIENCE

Face images of the same person can differ in expressions when smiling or

laughing. Also , some details such as dark glasses , beards , or moustaches can be

present.



CHPTER 3: FACE RECOGNITION USING EIGENFACES

This chapter deals with the complete explatation of face recognition using eigen-faces.

3.1Introduction:

Eigen-faces are a set of eigenvectors used in the computer vision problem of human face

recognition. The approach of using eigenfaces for recognition was developed by Sirovich and

Kirby (1987) and used by Matthew Turk and Alex Pentland in face classification. It is considered

the first successful example of facial recognition technology. These eigenvectors are derived

from the covariance matrix of the probability distribution of the high-dimensional vector

space of possible faces of human beings.

A set of eigenfaces can be generated by performing a mathematical process

called principal component analysis (PCA) on a large set of images depicting different human

faces. Informally, eigenfaces can be considered a set of "standardized face ingredients", derived

from statical analysis of many pictures of faces. Any human face can be considered to be a

combination of these standard faces. For example, one's face might be composed of the average

face plus 10% from eigenface 1, 55% from eigenface 2, and even -3% from eigenface 3.

Remarkably, it does not take many eigenfaces combined together to achieve a fair approximation

of most faces. Also, because a person's face is not recorded by a digital photograph, but instead

as just a list of values (one value for each eigenface in the database used), much less space is

taken for each person's face.

The eigenfaces that are created will appear as light and dark areas that are arranged in a specific

pattern. This pattern is how different features of a face are singled out to be evaluated and

scored. There will be a pattern to evaluate symmetry, if there is any style of facial hair, where the

hairline is, or evaluate the size of the nose or mouth. Other eigenfaces have patterns that are less

simple to identify, and the image of the eigenface may look very little like a face.



The technique used in creating eigenfaces and using them for recognition is also used outside of

facial recognition. This technique is also used for handwriting analysis, lip reading, voice

recognition, sign language /hand gesturesinterpretation and medical imaging analysis.

Therefore, some do not use the term eigenface, but prefer to use 'eigenimage'.

The task of facial recogniton is discriminating input signals (image data) into several

classes (persons). The input signals are highly noisy (e.g. the noise is caused by differing lighting

conditions, pose etc.), yet the input images are not completely random and in spite of their

differences there are patterns which occur in any input signal. Such patterns, which can be

observed in all signals could be - in the domain of facial recognition - the presence of some

objects (eyes, nose, mouth) in any face as well as relative distances between these objects. These

characteristic features are called eigenfaces in the facial recognition domain (or principal

components generally). They can be extracted out of original image data by means of a

mathematical tool called Principal Component Analysis (PCA).

By means of PCA one can transform each original image of the training set into a

corresponding eigenface. An important feature of PCA is that one can reconstruct reconstruct

any original image from the training set by combining the eigenfaces. Remember that eigenfaces

are nothing less than characteristic features of the faces. Therefore one could say that the original

face image can be reconstructed from eigenfaces if one adds up all the eigenfaces (features) in

the right proportion. Each eigenface represents only certain features of the face, which may or

may not be present in the original image. If the feature is present in the original image to a higher

degree, the share of the corresponding eigenface in the ”sum” of the eigenfaces should be

greater. If, contrary, the particular feature is not (or almost not) present in the original image,

then the corresponding eigenface should contribute a smaller (or not at all) part to the sum of

eigenfaces. So, in order to reconstruct the original image from the eigenfaces, one has to build a

kind of weighted sum of all eigenfaces. That is, the reconstructed original image is equal to a

sum of all eigenfaces, with each eigenface having a certain weight. This weight specifies, to what

degree the specific feature (eigenface) is present in the original image.

If one uses all the eigenfaces extracted from original images, one can reconstruct the

original images from the eigenfaces exactly. But one can also use only a part of the eigenfaces.

Then the reconstructed image is an approximation of the original image. However, one can



ensure that losses due to omitting some of the eigenfaces can be minimized. This happens by

choosing only the most important features (eigenfaces). Omission of eigenfaces is necessary due

to scarcity of computational resources.

How does this relate to facial recognition? The clue is that it is possible not only to

extract the face from eigenfaces given a set of weights, but also to go the opposite way. This

opposite way would be to extract the weights from eigenfaces and the face to be recognized.

These weights tell nothing less, as the amount by which the face in question differs from

”typical” faces represented by the eigenfaces. Therefore, using this weights one can determine

two important things:

1. Determine, if the image in question is a face at all. In the case the weights of the image

differ too much from the weights of face images (i.e. images, from which we know for

sure that they are faces), the image probably is not a face.

2. Similar faces (images) possess similar features (eigenfaces) to similar degrees (weights).

If one extracts weights from all the images available, the images could be grouped to

clusters. That is, all images having similar weights are likely to be similar faces.



3.2 Overview over the algorithm

Figure 1: High-level functioning principle of the eigenface-based facial recognition algorithm

The algorithm for the facial recognition using eigenfaces is basically described in figure 1. First,

the original images of the training set are transformed into a set of eigenfaces E. Afterwards, the

weights are calculated for each image of the training set and stored in the set W .

Upon observing an unknown image X, the weights are calculated for that particular image and

stored in the vector WX. Afterwards, WX is compared with the weights of images, of which one



knows for certain that they are faces (the weights of the training set W ). One way to do it would

be to regard each weight vector as a point in space and calculate an average distance D between

the weight vectors from WX and the weight vector of the unknown image WX (the Euclidean

distance described in appendix A would be a measure for that). If this average distance exceeds

some threshold value , then the weight vector of the unknown image WX lies too ”far apart”

from the weights of the faces. In this case, the unknown X is considered to not a face. Otherwise

(if X is actualy a face), its weight vector WX is stored for later classification. The optimal

threshold value has to be determined empirically.

3.3 Eigenvectors and eigenvalues

An eigenvector of a matrix is a vector such that, if multiplied with the matrix, the result is

always an integer multiple of that vector. This integer value is the corresponding eigenvalue of

the eigenvector. This relationship can be described by the equation M × u = × u, where u is an

eigenvector of the matrix M and is the corresponding eigenvalue.

Eigenvectors possess following properties:

They can be determined only for square matrices

There are n eigenvectors (and corresponding eigenvalues) in a n × n matrix.

All eigenvectors are perpendicular, i.e. at right angle with each other.



3.4.1 Calculation of eigenfaces with PCA

In this section, the original scheme for determination of the eigenfaces using PCA will be

presented. The algorithm described in scope of this paper is a variation of the one outlined here.



3.4.2 Face Recognition Using Eigenfaces

− The distance er is called distance within the face space ( difs )



3.4.2 Face Detection Using Eigenfaces



3.7MERITS AND DEMERITS

MERITS:

1. Euclidian distance is only a single parameter used to classify the image.

2. Complete face information is taken into account for recognition.

3. Relative insensitivity to small or gradual change in the face image.

4. Better in speed , simplicity and learning capability.

DEMERITS:

1. If lighting effects and the position of the face with respect to the camera is varied

Greately then accuracy will effect.

2. Only gray scale images can be detected.

3. A noisy image or partially occluded face causes recognition performance to

degrade gracefully.



3.8APPLICATION:

Face recognition system has following application:

Given a database of standard face images (say criminal mug shots), determine whether or

not a new shot of a person is in database.

Authorize users to allow login access.

Prepare a surveillance camera system residing at some public place which automatically

matches the input faces with criminal database and gives alert if the results are matched.

Match the person with his passport image, licence image etc.

Can be embedded and used for security purposes like mobile , laptop, palmtop.



Chapter 4.Some images of face recognition:



EIGENFACES

EIGENFACES RECONSTRUCTION

References:



T. M. Mitchell. Machine Learning. McGraw-Hill International Editions, 1997.

D. Pissarenko. Neural networks for financial time series prediction: Overview over recent research. BSc thesis, 2002.

L. I. Smith. A tutorial on principal components analysis, February 2002. URL http://www.cs.otago.ac.nz/cosc453/student_tutorials/principal_components.pdf. (URL accessed on November 27, 2002).

M. Turk and A. Pentland. Eigenfaces for recognition. Journal of Cognitive Neuroscience, 3 (1), 1991a. URL http://www.cs.ucsb.edu/ mturk/Papers/jcn.pdf. (URL accessed on November 27, 2002).

M. A. Turk and A. P. Pentland. Face recognition using eigenfaces. In Proc. of Computer Vision and Pattern Recognition, pages 586-591. IEEE, June 1991b. URLhttp://www.cs.wisc.edu/ dyer/cs540/handouts/mturk-CVPR91.pdf. (URL accessed on November 27, 2002).


Seminar Report

Documents

Transcript of Seminar Report