Man Made Security System Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Researched area with related subject: This research paper related with security system and show relationship of bio metric and dsp techniques. This study work has been done by us since from 12th march 2009 to 1st February 2010.we study many books and papers related biometric and dsp.finally we have selected face recognition with modification of the password system which makes the system most robust following outline with description define each and every term of that project related its mathematics platform as well as real time result through programming.

Introduction to security system

Abstract of system

Mathematically proof of the system

Dsp based techniques used in the system

Importance of the Eigen face and its calculations

Mathematics of Eigen face for face recognition

Hardware and software descriptions

Real times results

Future aspects


Introduction to security system :

Security system is one major task nowadays but electronics technologies through its application makes the security system most reliable .there are many software based security system available but each system have different principles and algorithms required. Everyone should need high security and surveillance system. It meant which give 100% authentication results. For more authentications system the biometric based electronics system played life time and most reliable role. Biometric security system in terms of electronics technology defined as "The system through which living person are recognizing or identifying by automatic system according to considerations of the physiological and behavioral characteristics". Whenever biometric based system is required than we need system with its supporting tools which mathematical as well as programming results. For consideration of that purpose DSP("Digital Signal Processing")play a vital role because DSP Provides easy techniques for mathematics platform as well as programming platform. some biometrics system are enlist below

Signature identification

Hand geometry

Body Oder identification

Iris identification

Retinal identification

Abstract of system

There are many parts of the human body ,which are used in biometric project for recognitions purpose of the highly security system. Face is one of major and secure part for identification. we have done the software based research project for face recognition. This system based on the password and it take password from user and captures the picture and verify it according to programming and physical based appreance conditions and in true entry it gives access grant, in case of false entry it denied. There are many techniques of doing this but we have selected PCA (Principles component Analysis) through applicable of Eigen Faces approach because PCA give more suitable dimension of the dataset.

Mathematically proof of the system

In biometric can never worked without its mathematics background. we use matlab as programming tool (MATLAB VERSION 7) and it is very important to proof all the mathematically terminology through programming. This research project based on the following mathematics factors

Standard deviation


Co variance

Eigen vectors and Eigen values

Standard deviation:

Standard deviation is the factor which gives the measurements of the spread of the data from the data set. mathematically standard deviation defined as "mean of the data from average distance towards setpoint"formulatically define as below


s=standard deviation

x(bar)=mean set of X

NOTE: Due to standard deviation we can get mean distance of image of the more average image (average mean) from the images sets (database) which makes more authentication of the images.


Variance is the square of the standard deviation formulaically defined below

Both variance and standard deviation mathematically used to measure spread of the data of the data set.

NOTE: Due to variance we get more demsionality of the images of the images set

Co variance :

Both the standard deviation and variance are used to measure the one dimension of the given image or target to measure the three dimensions. The mathematically factor used to measure more than one dimension is known co variance.formulatically defined below

In co variance between X and Y represented as below

NOTE: more than one dimension co variance give following result

If expected values (X and Y) are positive than our dimension of the data set increase positively

If expected values are negative than our dimension of the data set increase negatively

If expected values (X and Y) are zero than our dimension of the data set became independent

Eigen vectors and Eigen values :

Eigen vectors and Eigen values are used to represent the data (images) in the matrix form and give final result with suitable dimension. Below relationship define Eigen vectors and Eigen values.

M Ã- u = LAMBDA Ã- u,


u =An Eigen vector of the M matrix

LAMBDA =Eigen values of the Eigen vector

Properties of the Eigen vectors and Eigen values:

Eigen vectors and Eigen values are used to form square matrix and square matrix give suitable dimension of the mean image

Eigen vectors are at right angle with each other

Dsp based techniques used in the system:

There are many techniques with most suitable applications used for recognition the human parts. We are going to recognition the face. We have worked through PCA (Principle component techniques).following question come in our mind when we studied this technique.

What is PCA?

How can we find PCA for Face Recognition?

How can we apply PCA on the face recognition?

Why we select PCA Technique?

What is PCA?

It is mathematical technique, which used to defined the simplicity of the data set.PCA reduce the dimension of the data set according to verification of the standard deviation, variance , co variance,eigen vector and Eigen values' with summarized words PCA have following functions

Reduce the dimension of the database

Recognition the meaningful images or data according to verification of mathematics factors

How can we find PCA for Face Recognition?

Finding of the PCA depend upon two factors

Normalization of the data set

Evaluation of the Eigen vectors and Eigen values

Normalization of the data set:

Before study the normalization process, we need to suitable dimension of the images (data), regarding this co variance matrix is used for that purpose .our task is to get mean centering of the image, so co related the images and get one most suitable co variances matrix .it can done by using normalization because normalization is the statistical process of the spread of the data at suitable range. If we don't consider normalization it became difficult for us to manage the various components and get one suitable component that is called PCA

Evaluation of the Eigen vectors and Eigen values:

Actually the PCA depend upon the co variance matrix. It is collection of Eigen vectors and Eigen values ,select according to the dimension based and give priority to the suitable dimension. Highest priority component is known 1st [principle component and next highest one is known 2nd principle component and so on.

It is easy to identify the 1st ,2nd PCA component………………………………nth component but we need to verify it. The best way to verify it by using some mathematics techniques. Consider below example.










M= = 10* + 10*


From above given matrix, our task is to defined Eigen vectors and Eigen values .according to statistical mathematics the 1st component is Eigen vector and 2nd Eigen values and so on

How can we apply PCA on the face recognition?

Now we know about PCA but our task is not just knowing about PCA but its applicable on our recognition system (face recognition).applying of the PCA on face recognition depend upon following factors.

Make training sets of the images or data(

Arrange the training set in particular form, so that training sets are known data set or image space

Apply the PCA and intend it to work in the 1st phase (Normalization) because making good dimensions

Apply the PCA for 2nd phase (Eigen vectors and Eigen values)

Get linear combination of the Eigen vectors and normalization for result purpose

Why we select PCA Technique?

There are many techniques but PCA is one of the most suitable for mathematically as well as programming level and in the result also and most important reason of using PCA defined below.

Easy to make programming of the PCA and their mathematics conditions

Reduce the dimension with wide range

Easy to applicable for face recognition through verifying mathematics factors

Importance of the Eigen face and its calculations:

In this face recognition project Eigen face approached based on the principal component analysis and which follow the terminologies of the Eigen vectors and Eigen values. the Eigen face approach based on the PCA for given high quality image result from Eigen face space(data base) .proper dimension of the mean image depend upon proper projection of the Eigen vectors and Eigen values. The following are key factors which have done by applicable of Eigen face approach.

Input image with arrangement of suitable dimension in face space

Normalization of the face space

User password system

Final result recognition process

Mathematics of Eigen faces for face recognition:

Face recognition with application of the Eigen face depend upon following factors with respect to its mathematics

Initialization of the capture image

Mathematics of the key vector

Mathematics of the face space

Mathematics of resulting face

Initialization of the capture image:

Our first task will to make proper arrangement of the input images mathematics is necessary for software point view. Our first step is formation of the co variance matrix and 2nd step is comparing with existing images in the data base. Consider is new image and n images vector formation occur with dimension of Nx1.and note mean image (capture's image) will subtract from average image

Mathematics of the key vector :

The resulting image in the form of vector and give average image (average image -capture image=mean image) and mathematical define below

Mathematics of the face space :

The arrangement of the images in particular form is known face space. Mathematics of the face space in the form of the matrix and we give elements of the matrix as class in place of rows and columns. Such as

Class #01 ________________________________w1

Class #02 ________________________________w2

Class #03 ________________________________w3

Class #04 ________________________________w4

Up to

Class #nth ________________________________Mth

Mean result will be

Mathematics of resulting face:

Resulting image is formed according to following mathematics

Input image mean image

According to this mathematics there are four ways to recognitions the image

face space and a face class both have same co variance matrix (recognized correct)

face space but face class both have not same co variance matrix ( non recognized)

face space and face class have different face key vectors (unknown a face )

face space and face class have not different face key vectors (unknown a face )

Hardware and software descriptions:

Basic diagram of the face recognition shown below

Input Image

Normalized image

Face space

Face ID


It is PC based and only two hardware devices are connected with system. One is camera for capture image and one external circuit for access grant and access Denied.

Parallel port interfacing circuit


Software is the Matlab,which used as programming tool and system work according to below graphical algorithm.

Real times results:

Real time result contains on following blocks according to interconnecting with software code

Data base

Access grant

Access denied

Data base

Access Granted

Access Denied

Future aspects or application:

It is the one easy software based project and used in the offices ,colleges ,universities for high security authentications

Matlab is used as programming tool due to this we can use just in colleges and universities because matlab have not suffient effiency to give fast result for comparing large amount of data base( contain not more than 200 classes and each class contain with 5 images).for its fast speed for given result of large data base c++ will be better platform and our work is in progress for that purpose.

In future through this software image can capture from live video and recognized. After such nice capturing and recognitions technique than software will even become more authenticated and robust for security purpose.

Due to password this system is most reliable and secure