Dr. A.P.J.

ABDUL KALAM TECHNICAL UNIVERSITY LUCKNOW

(UTTAR PRADESH)

A
SYNOPSIS ON
“ATTENDENCE MANAGEMENT SYSTEM USING
FACE RECOGNITION”
SESSION 2019-2020

FU
INSTITUTE OF ENGINEERING & TECHNOLOGY BAREILLY

TU
In partial fulfillment for the award of the degree of
BACHELOR OF TECHNOLOGY
(COMPUTER SCIENCE AND ENGINEERING)
2019-2020

SUBMITTED TO :-
D. R. K. P. GANGWAR (H. O. D.)
RE SUBMITTED BY :-
ANKIT VERMA (1647610015)
AMBRISH SHARMA HIMANSHU SHARMA (1647610030)
PANKAJ KUMAR (1647610052)
SHUBHANESH (1647610076)
 Improve your memory by doing unforgettable things.

DECLARATION

I declare that the project work with the title “ATTENDANCE MANAGEMENT
SYSTEM USING FACE RECOGNITION” is my original work done under Mr.
Dr. K. P. Gangwar and Mr. Ambrish Sharma, Future Institute of Engineering and
Technology Bareilly. I have learned and followed all the rules and regulations
provided by the Institute while writing this synopsis report. This project work is being
submitted in the fulfillment of the requirements for the degree of Bachelor of
Technology in Computer Science and Engineering at Future Institute of Engineering
and Technology for the academic session 2019-2020.

Ankit Verma (1647610015)

Himanshu Sharma (1647610030)
Pankaj Kumar (1647610052)
Shubhanesh (1647610076)

2
 CERTIFICATE

This is to certify that

SHAP group (Ankit, Himanshu, Pankaj, Shubhanesh)

Of B.Tech. Computer Science and Engineering s have successfully submitted synopsis report on
“ATTENDANCE MANAGEMENT SYSTEM USING FACE
RECOGNITION“, in partial fulfillment of the Degree of Bachelor of Technology in Computer
Science and Engineering under the guidance of “Ms. SAMIKSHA BISTH” and “Mr. DEEPAK
KUMAR”, from Future Institute of Engineering and Technology Bareilly, U.P. in the year 2019-
2020.

Project Guide
AMBRISH SHARMA

HOD, CSE DEPARTMENT

 Dr. K.P. GANGWAR

Internal Examiner External Examiner

DATE:
3
 CONTENTS

 Chapter 1
 ABSTRACT
 INTRODUCTION
 Flow chart
 Chapter 2
 Face detection
 Introduction
 Segmentation based on color
I. RGB Color Space
II. HSV Color Model
III. YCbCr Color Model
 PROPOSED APPROACH

 Chapter 3
 Face recognition
 Introduction
 Principal Component Analysis (PCA
 Steps Used in PCA Algorithm
 Mathematical Approach
 EXPERIMENTAL ANALYSIS
 Experimental Results

 Chapter 4
 Attendance Registering
 Introduction
 Creating Attendance Sheet
 Marking Attendance of the Student

 Chapter 5
 Conclusion
 Future Work
 Chapter 1
Face Detection
5
 ABSTRACT

Uniqueness or individuality of an individual is his face.

In this project face of an individual is used for the purpose of attendance making
automatically. Attendance of the student is very important for every college, universities
and school. Conventional methodology for taking attendance is by calling the name or
roll number of the student and the attendance is recorded. Time consumption for this
purpose is an important point of concern.
Assume that the duration for one subject is around 60 minutes or 1 hour & to record
attendance takes 5 to 10 minutes. For every tutor this is consumption of time. To stay
away from these losses, an automatic process is used in this project which is based
on image processing.

In this project face detection and face recognition is used. Face detection is used to locate
the position of face region and face recognition is used for marking the understudy’s
attendance. The database of all the students in the class is stored and when the face of the
individual student matches with one of the faces stored in the database then the
attendance is recorded.
6
 INTRODUCTION

Attendance is prime important for both the teacher and student of an educational
organization. So it is very important to keep record of the attendance. The problem arises
when we think about the traditional process of taking attendance in class room. Calling
name or roll number of the student for attendance is not only a problem of time
consumption but also it needs energy. So an automatic attendance system can solve all
above problems.

There are some automatic attendances making system which are currently used by much
institution. One of such system is biometric technique. Although it is automatic and a step
ahead of traditional method it fails to meet the time constraint. The student has to wait in
queue for giving attendance, which is time taking.

This project introduces an involuntary attendance marking system, devoid of any kind
of interference with the normal teaching procedure. The system can be also
implemented during exam sessions or in other teaching activities where attendance is
highly essential. This system eliminates classical student identification such as calling
name of the student, or checking respective identification cards of the student, which
can not only interfere with the ongoing teaching process, but also can be stressful for
students during examination sessions.
7
FLOW CHART
8
 Chapter 2
Face Detection
9
 INTRODUCTION

Face detection is defined as finding the position of the face of an individual. In other
word it can be defined as locating the face region in an image. After detecting the face of
human its facial features is extracted and has wide range of application like facial
expression recognition, face recognition, observation systems, human PC interface and so
forth. Detecting face in an image of single person is easy but when we consider a group
image of an image containing multiple faces, the task becomes difficult.

For the application of face recognition, detection of face is very important and the
first step. After detecting face the face recognition algorithm can only be functional.
Face detection itself involves some complexities for example surroundings, postures,
enlightenment etc.

There are some existing methodologies for detection of face. Some of them are skin color
based, characteristic or feature based (feature like mouth, nose and eyes) and neural
network based. Among the above techniques, the skin based procedure is well thought-
out as simplest one. The approach premeditated and applied in this thesis is the skin color
based face detection method. The algorithm is pretty dynamic as numerous people face
can be detected at one time from an image containing many people. In this project YCbCr
color model is used to detect the skin of human being
10
 Segmentation based on color

The Segmentation can be defined as the conception of subdividing a given image into its
constituent region. Segmentation based on the color of skin is picking up its supremacy
in current time. Skin based segmentation is being studied for the reason that of its
dynamic research in content- based picture illustration.
In the case of face detection, segmentation is used to find locate the face boundary of
face region in an image. Once the face region is found, we can apply various processing
like image editing, various coding, and image indexing and client intuitiveness intention.
Moreover, face detection is the first step required for the purpose of recognition of face
and its expression using various processes.
Color of Skin of an individual depends on various biochemical components like the
melanin content, pigmentation of skin and much more. The skin color is belongs to
certain range in the total color space. Consideration should be taken into account that the
skin should not be abnormal. Used algorithm in this project takes the benefit of face
color association to limit the face search to areas of an input image that have at least the
accurate color components. There are many existing algorithm for segmentation but used
algorithm is the simplest one.

I. RGB Color Space

rel
ate
Red, green, blue are the three color component that constitute the RGB color space.
d
These red, green, and blue colors combine in a definite proportion to yield a different
to
color. RGB color model can be represented by a 3-dimentional cube with the three colors
at the corner of the cube and in each axis as shown in Figure.
At the origin of this cube black color is present. White color is present at the opposite
corner of the 3-dimesional cube i.e., at the opposite diagonal of the cube. Gray color scale
is represented by the line from black color at the origin to the white color at opposite
corner. Red is (255, 0, 0) when we consider a 24-bit color graphics system with a color
bit of 8-bit per color channel. When we consider the 3-dimensional color cube, red is at
(1, 0, 0).
This model simplifies the design of the pc graphics frameworks yet is not perfect for all
13 type of application applications. These three color component are very closely
each other, which makes it hard to implement some of the algorithm for image
processing.

11
II. HSV Color Model

There are some problems which are associated with the RGB color space model. One
important issue connected with RGB is that it doesn’t consider the effect of light on the
color of skin, which generally cause some wrong in turn. This problem can be solved by
using HSV model for skin color. Here H stand for hue which is depth of color, S stand for
saturation or the purity of color and V stand for the value of intensity of light or the
brightness of color. The model is shown in figure.
The red, blue and yellow color is represented by hue and which has the range from 0 to
3600. While using HSV color space we don’t need the information of the proportion of blue
and green color needed to produce a different color. Only thing needed is that by altering
the hue we can get the desired color of our interest. The purity of the color is represented by
the saturation value, which has the range of 0 to 100%. Let’s take an example of getting
pink color from that of the dark red color, which can be done by simply modifying the
saturation value .value gives idea about the brightness of the given color and it gives the
12

achromatic thought of the given color. The range of value extend

from zero to hundred. ‘H’ and ‘S’ gives the required information about the color of skin.
The pixel value of skin color should satisfy the below condition in order to consider that
as skin.

“0 <= H <= 0.25” and “0.15 <= S <= 0.9;”

III. YCbCr Color Model

This color model has more advantage than the upper two model as discussed above and it
uses the chrominance value to extract the skin color region of a given image. ‘YCbCr’ or
‘Y’CbCr’ color space are generally used in the digital image processing. Y is the
luminance, luma component is represented by Y’ while Cb and Cr are the blue difference
and red difference of the chroma component respectively. YCbCr is not a real color
13
space. This is just another way of encoding the RGB color space. The YCbCr values can
only be obtained only if the original RGB information of the image is available.

14
 PROPOSED APPROACH

Here I have described the algorithm that I have used for the purpose of face detection.
The algorithm is the combination of RGB and HSV algorithm. The efficiency of the
proposed algorithm is more when single face in a given image is taken into consideration.
In case of no of face more than one in an image, the acquired input image may result in
detection of some false face. The steps of the algorithm is discussed as follows 17
STEP 1: (Acquisition of the input image)
The input image can be acquired using a variety
of conduct for example we can take the image by using camera directly or we can take
the image from the stored database folder. The second way is more effectual as the
debugging process becomes much faster.

STEP 2: (Extraction of the skin type using HSV color space)

In this algorithm the
HSV color model is used which consider the effect of light. In MATLAB ‘rgb2hsv’
command is used to get the HSV image from that of the RGB image. Later than that each
of the three H, S, V component is extracted and the extracted image gives information for
the color of skin. A pixel whose H and S component satisfies the equation.
0<=H<=0.25; 0.15<=S<=0.9;
Is considered as a pixel in the individual. In Figure 5 the HSV color space image
is shown and in figure 7 the extracted skin color image is shown.

15
STEP 3: (output image is transferred to the rgb color and produce binary image
based on the value of R, G and B)
After getting the skin color region the image is
converted into the RGB image i.e, only the skin color region is converted. Next the
R,G,B value is extracted by assigning the value of the three channels into three variables.
If the R,G,B value of the pixel satisfies the below equation then it will classified as skin.

If the pixel satisfied condition will become white else it will be of black color as
shown in the Figure.

STEP4: filling of black dot inside the white region.

The converted binary image contains black spot, and that can create problem in next
processing. To fill the black spot I used the ‘imfill’ command of MATLAB. The resulting
16

image is shown in figure 7.

STEP 5: morphological operation is done to remove the small areas which are very
small to be a face.
This step is the most critical and additionally the best module in the
entire face identification process. As expressed before, the algorithm based on skin color,
if other exposed part of the body is more than the pixel value of the face region than it
will create problem. Figure 8 shows the small exposed body part i.e, other than face is
removed.

Step6: (drawing bounding box around each face region)

“bwboundaries” command is
used to get the boundary around the detected face. In the below figure it is shown
with green color rectangular boundary.
17
 Chapter 3
Face recognition

18
 INTRODUCTION

The recognition of face of human is challenging in computer-human interaction. The face

is our essential center of consideration in societal life playing a critical part in assigning
identification and emotion of the person.
We can perceive various appearances adapted all through our lifespan and distinguish faces
initially even following quite a while of detachment. This expertise is very vigorous
notwithstanding of substantial varieties in visual boost because of evolving condition,
maturing and diversions, for example, facial hair, glasses or changes in haircut.
Computational models of face acknowledgment are fascinating in light of the fact that
they can contribute to hypothetical learning as well as to functional applications. PCs that
identify and recognizes the face could be connected to a broad assortment of
undertakings together with criminal recognizable proof, security framework, image and
film handling, identity confirmation and human-PC interaction.
Tragically, adding to a computational model of face recognition and acknowledgment is
very troublesome in light of the fact that faces are perplexing, multidimensional and
important visual stimuli. This project uses the principal component analysis (PCA) for
face recognition. There are many other face recognition algorithm, but principal
component analysis (PCA) based face recognition is the simplest one for face
recognition.

Principal Component Analysis (PCA)

PCA was invented by Karl Pearson in 1901. It involves the mathematical method to
transforms an amount of probably interrelated variables into a numeral of
unorganized variables is called the principal components, related to original variables
by an orthogonal transformation.
The transformation can be defined as that the first principal component has as high
variance as it is possible and for every succeeding component thus has the highest
variance conceivable under the imperative that it be orthogonal to the former
components.
19
 Steps Used in PCA Algorithm

Step1:
Acquire training set of ‘N’ number of images at the initial stage. In this project the
images are of 92*112 pixels each. Training set is shown in the figure.
Step2:
Calculation of the eigenface from the “N” training set images keeping only few M images
that is correspond to that of the highest Eigen values. The M images describe the “face
space”. When new faces encountered, the “eigenfaces” can be recalculated accordingly.
Step3:
The corresponding distribution of the “M” dimensional weight space for every known
individual is calculated by projecting their respective face images onto “face space”.
Step 4:
Compute set of the weights anticipating or projecting the data picture or input image to
M “eigen faces”.
Step5:
Determine if the given image is face image or not by checking to the closeness of
given image or picture to “face space”. 27
Step 6:
If the image is sufficiently close enough, then classify the weight pattern as either an
unknown or as a known person based on measured Euclidean distance.
Step7:
If the image is sufficiently close enough then refer to the recognition is successful and
give applicable information about recognized face from the database which hold data of
faces.
20
21
 Recognizing an unknown face

Mathematical Approach

Let I1, I2 Im be the ‘M’ set of train images.

The Average face of set can be defined as
‘A’ = (1/M)*∑ ( ) =1
Individual face difference from the average face is given by
Ф = -A 29
When it is subjected to principal component analysis (PCA), the large set of vectors
seeks a set of M orthogonal vectors Un, which best describes the distribution of data.
22

The kth vector Uk is chosen such that

= (1/M) ∑ [( ) ∗Ф ] =
is maximum, subject to
( ) = ={ , = ,
The vector and the scalar are the eigen vectors and eigen values respectively of the
covariance matrix
C= (1/M) ∑ (Ф) =1 (Ф)’
C= B*B’
Where the matrix
B= [Ф1Ф2….. Ф ]

EXPERIMENTAL ANALYSIS

The database or training set of student is created. The data base contain photo of 10
students. Each student having ten images so the total image is equals to 90. The
training set is shown in figure. Each image is of 92×112 pixel.
The test image set consists of ten images of the nine students. The image in the test set is
different i.e. they are not taken from the database.
The test set can be shown in figure as follows.

23
 Experimental Results

When an input image is given, which is not present in the data base, the input image is
processed and the best matching i.e, where the elucidance distance is less that image is
considered as the recognized image. The experimental result is shown in figure. Where
the recognized image is 16th image from the database and the recognized image is
himanshu, which is shown in the command window of the MATLAB.

24
 Chapter 4
Attendance Registering

25
 INTRODUCTION

Class attendance is very important aspects for the students studying in the colleges or
schools. For an organization to be successful, it needs precise and quick method for
recording the performance of the individuals inside this organization. Attendance gives
the data of the individual whether that particular person is physically present or absent.
The traditional method of calling roll number or name of the student for marking
attendance is time consuming or wastage of time during the class hour.
In this project an automated student attendance system is made, which is based on face
recognition. The recognized face of the individual is used for marking the attendance.

Creating Attendance Sheet

As the database contain images of nine students, so the attendance sheet should
contain name of the nine students. The 9×1 cell array is created which contain the
name of the nine student as shown in the figure.

26
 Marking Attendance of the Student

When the face of the particular student is recognized, attendance is marked for that
student for that particular day. If some students are absent then no attendance is marked
for them. For attendance marking logic 1 is written. In the 9×1 cell each cell contain
name of the student, and attendance is marked for the student shown in figure15.
The attendance sheet is stored in the variable name data. Consider the second name
stored in the cell array i.e, himanshu. When himanshu is recognized for let’s say five
times i.e, one in each day then attendance is marked five times for that particular
student. The particular case is shown in figure 16. Like this other student attendance is
marked according to their presence in that particular day.

In Figure 16 first row showing 1×3 cell, which means the particular student is present for
2 days. The first column is used for storing the name of the student and so the next two
column is used for marking the attendance.
Similarly the same is applicable for other raw of the 9×1 cell. Take another example of
raw no 2, which is showing 2×5 cell, which means that particular student is present for 4
days. The marked attendance of the student in the second raw cell is shown in Figure 17.
27
28
Chapter 5
Conclusion

29
 Conclusion

The face detection and recognition algorithms were studied thoroughly taking number of
the test from different varying condition images. For face detection combination of RGB
and HSV model algorithm is used.
For face recognition principal component analysis method is used. Attendance of the
student are marked using the recognized face of every individual student and the data
is stored in an attendance sheet.
The attendance of every student marked automatically by recognizing their face with
the face present in the data base.

Future Work

Further work can be done on this project to alert the student by sending SMS regarding
the attendance. For this purpose GSM module can be used. SMS alert can be given to
the parent of the student.

30