Professional Documents
Culture Documents
Abhishek Report
Abhishek Report
BACHELOR OF ENGINEERING
IN
ELECTRONICS AND COMMUNICATION ENGINEERING
Submitted by:
ABHISHEK G.S
[1CG18EC400]
2020-21
Channabasaveshwara Institute of Technology
(Affiliated to VTU, Belagavi & Approved by AICTE, New Delhi)
(NAAC Accredited & ISO 9001:2015 Certified Institution)
NH 206 (B.H. Road), Gubbi, Tumkur – 572216. Karnataka.
CERTIFICATE
UNDERTAKING
BONAFIDE CERTIFICATE
Guide
A GREAT DEAL OF TIME AND LOT OF EFFORTS have gone into completing
this technical seminar and documenting it. The number of hours spent in getting through
various books and other on-line materials related to this topic chosen by me have reaffirmed
Several special people have contributed significantly to this effort. First of all, I am
I, acknowledge and express my sincere thanks to our beloved Director & Principal,
Dr. Suresh D S, for his many valuable suggestion and continued encouragement by
I, express my sincere gratitude to Dr. Nagananda M.Tech, PhD, Professor and Head,
guidance, support and suggestions throughout the period of this technical seminar.
Finally, I would like to express my sincere thanks to all our teaching and non-
teaching staff; they were extremely helpful in assisting me in using the excellent facilities
of the institute.
ABSTRACT:
Nowadays face recognition is important in cities for development of biometric
applications. It requires many human resources than different types of security measures,like
fingerprinting attendance is compulsory to lecturer for the presence of students.Face
recognition is developing technology it rapidly growing to safe security. This technology
improved a lot in recent years and give a good performance.It is a way of identifying persons
identity by using their face.In otherwords it is also called as biometrics and face recognition is
used to preventing Id fraud and identifying theft and also used for issuing documents identity.It
is a computer technology used for many applications to identify human faces .It has more useful
than biometrics like palmprint.This system is used for biometrics to match facial features from
photo or video.It also compares the required information to find a match.It has ability to detect
first and later it recognizes the face and it has ability to find huge number of people.In many
situations facial recognition system can be safety and security.Todays critical pandemic
situation has transformed the way of educating a student.Education is accept through online
platforms like online courses and online teaching and it has also changed the way of
assessments through online .Many educational institutions adopting a online examination
portals for the assessments of the students.Attendance plays a vital role in online education it
shows the students sincerity towards studies.So biometrics is used for students, engineers as
well as scientists this face recognition system. In many countries people used this system for
business and other educational systems used this system for conducting online exams.This
paper presents the how face recognition system helps in many fields and for the conduction of
online exams we need biometrics to find a individual's face and finger print.So in recent days
biometrics is more important for many educational institutions. It is considered as more populer
resarch about recognition of face of human beings.nowadays in schools and colleges attendence
is more important it is considered as identity of a individual person.Many cameras is used to
detect the face so it will very safe and it has huge amount of data and it is very fast to detect
face so it is used.so attendence system is used with the help camera's so it will be used for many
companies for identifying the person's attendence and it will be stored for many time so
attendence system by using cameras are very helpful to access the data and more valuable
time.Biometrics are used to understand the things and development of present technologies and
trends.This biometrics has more huge applications withrespect to consumption of time and
more security for the customers.The resarchere studies in many fields for the importantance of
face recognition system so it will be more usefull for in all technologies development and all
fields.face recognition is helps in the field of vision so by seeing face with the help of camera
we can easily findout the persons human face.It is very useful for security purposes in different
large scale inustries so identify the person life.This system is helpful for many fields but it is
very difficult to detect to face recognition for system.Face recognition is very helpful than other
techniques.so it is very useful in all the needy purposes but nowadays many terrorist and
criminals are covered their face so with the help of masked face recognition we can identify
their face so it is very usefull to identify the thefts,criminals and many terrorists so this systems
are very important and also it is very complicated to identify human face with the help of
systems but humans can identify so we need to make system should also identify from the
images so it is very helpfull even in all organization sector and public faces to avoid many
illegal activities so face recognition system is used in all fields.
CONTENTS
CHAPTER -1
1.1Introduction ...................................................................................................... 5
CHAPTER -2
2. 1 LITERATURE REVIEW ..................................................................................................7
CHAPTER -3
3.1METHODOLOG……………………………………………………………………………9
CHAPTER -4
4.1RECOGNITION……………………………………………………………………………11
CHAPTER -5
5.5 DESIGN PROCESS ………………………………………………………………………17
CHAPTER -6
6.6RESULTANDDISCUSSION………………………………………………………………19
CHAPTER -7
7.1FACERECOGNITIONPROCESS………………………………………………………...22
CHAPTER -8
8.1FUTUREWORK……………………………………………………………………………24
CONCLUSION-9………………………………………………………………………………25
REFERENCES-10……………………………………………………………………………..26
LIST OF FIGURES
CHAPTER - 1
1.1 Introduction
Attendance system in the world of education is an important thing in a lecture as proof
of the presence of students. In one of the educational institutions that the author analyzes, the
presence of students determines whether a student has the right or not to take a lecture Final
Examination (UAP) with a percentage of 80% of 13 meetings in one period. Attendance
activities that have been running so far have used the system but there are still weaknesses,
namely accuracy in ensuring that the student who does it himself or other parties who do
attendance and the teacher only looks at the student login activities.
Biometrics is a developing technology, which has been widely used in forensics, safe
access and prison security. Biometric systems are basically a pattern recognition system that
recognizes someone by determining authentication by using different biological features,
namely fingerprints, retinal scanning, iris scanning, hand geometry, and facial recognition
that causes physiological biometrics and behavioral characteristics are speech recognition,
keystroke - scan, and signature scan.
At this time the exam system that takes place in the Binus online runs manually, where
students come to several branches of Binus that are near their place, then carry out the test
according to the time specified. it is not efficient because more and more online students are
scattered in several regions and several Binus branches that have not been able to cover it.
The aim to be achieved in this study is to make a prototype of an online exam application
design in order to recognize students who will take the exam with face recognition methods.
As for the expected results in the design of this exam application prototype:
The benefits generated in this study are expected to overcome the weaknesses of student
absenteeism which currently runs manually on Binus Online and prevents misuse of the
existing attendance system, for example occurs with attendance manipulation by entrusting
accounts / logins to other students or other people.
CHAPTER – 2
Face verification is a process of recognizing and matching faces. The use of biometrics
for recognition systems has the aim of increasing human comfort and security in the scope of
personal privacy and in a wider scope such as for an agency, the advantages of biometrics
have many benefits and advantages compared to traditional systems such as: manual signing,
use of passwords, PINs, cards and the key that has been applied to: entrance access,
attendance, ATM machines and others. According to Syed Navaz & Mazumder Face
identification systems also have advantages such as: Accurate, Cost-effective, non-invasive,
using legacy data, the only biometrics suitable for use and made as a backup mechanism.
Eigenface is the name given to a set of eigenvectors when used in face recognition in
the field of computer vision. In Layman's terms, Eigenface is a set of standardized face
ingredients from statistical analysis of many facial images.
The Eigen faces approach to facial recognition was developed by Sirovich and Kirby (1987)
The face recognition algorithm with the Eigenface method is carried out through several
stages, namely [8]:
Arrange the Flat vector Image Matrix
Calculating Middle Values or Mean (w)
Calculating the Difference between Training Images with Middle Values or Mean (w)
Calculating the Covariance Matrix Value
Calculating the Eigenvalue Value and Eigenvector
Finding the Eigenface Value
Identification Process
CHAPTER – 3
3. 1 METHODOLOGY
Figure 3.1 demonstrate our proposed method to the problem of robust face recognition under mask
and non-mask conditions. For face recognition at first detect the face portion from an image, find out the
face features and finally train a system for face recognition. During the training phase, for detecting face
portion we used Viola-Jones algorithm in our work and then PCA is used to efficiently represent the
gallery images. In this way we obtain a PCA feature space which is called Eigen faces or ghost face. In the
test case, a target face image (which can be masked or not) is given to be recognized. At first its PCA
representation is computed. As a result, potential facial components are identified. Then, the Eigen faces
are selected and used for recognition. The recognition is carried out by comparing the selected features
from the target image against selected features from the corresponding template images.
The input image is read and is converted into grayscale image of 8 bits. The image is
flipped along the vertical axis and is resized using Bicubic Interpolation over 4x4 pixel
neighborhood method. Bicubic Interpolation is used to find the position of two nearby points
and getting the orientation or gradients at those positions, which gets connected by a piece wise
cubic spline to get a smoother and image with a fewer interpolation artefacts.
It is also used to calculate the positions of the points between the two end points and
there by getting the new resized pixels value. For the facial recognition it is necessary to
enhance the facial features. The method used here is thresholding using Tozero, which is
performed on to the image where the pixels at the thresholding value is converted to black and
the other to subsequent lighter shades. Morphological transformation is performed on the image
in order to remove the noise and isolate individual elements.
Face detection is done using Haar Cascade Method (feature- based) which uses
machine learning algorithms to classify images into two categories- images containing faces
and images of objects. In this method we use cv2.CascadeClassifier() to classify the input
images. Further, feature wise extraction is performed on the images. Features are selected in
such a way that it could effectively classify images. For instance, the area of the eyes is
compared to the areas near the nose. As the region near the eyes are lighter than that of
region of the nose, this can classify the samples .
All such features are applied on all training images. And for each feature, there is a
threshold value, which identifies the face and non face images and classify them
accordingly. Since there is high chance for errors, we consider the features with minimum
error rate. During each classification, the misclassified images are given more weight and
the process is repeated. New error rates and weights are calculated until the requirements are
met. To ease the process, the features are grouped into different sets of classifiers and then
applied to the window one at a time . If a window fails the first stage, it is discarded. If it
succeeds, then the next set of features are applied. If a window passes all the stages, it is
considered as a face region.
CHAPTER – 4
4.1 RECOGNITION
Image recognition is done using Local Binary Pattern Histogram (LBPH) in which a
3x3 matrix window is moved along the image and the pixel value at the central location of the
matrix is calculated. The pixel value at the center is compared with the neighboring pixel
values. If the values measured are greater than or equal to the central pixel value then that
pixel value is read as 1 and the rest as 0. The values are read in clockwise order, forming a
binary value which is the local binary pattern of that specific area. This is done to the entire
image and LBPs are obtained.
The obtained LBPs are converted to its corresponding binary values and histograms are
plotted against each value. Each face will have a histogram in the training dataset. When a
new face is detected, its histogram is compared with the training dataset and the best matching
histogram is returned with the label of the person it belongs to. After that, when a face is
captured on the camera, the recognizer tries to identify the face and make a prediction. It
checks if the face data is already available in the training dataset.
If the data is available, it returns the histogram ID and an index. A portion of the face is
taken as a parameter by the recognizer. predict () and it will output the ID of the owner and
the accuracy of the recognizer. The confident index will output zero if it is a perfect match.
When a successful prediction is made, the probable id and the probability of a right match is
displayed over the image. Otherwise, an unknown label is displayed .
For the purpose of using the single snap mode, the backend consists of a function, snap()
This function is used to capture the face of a student, identify him/her and if identified, marks
him/her as present in the excel sheet of attendances. It consists of several variables which are
initialized first, and also contains a list of names of the trained faces. Then it initializes and
starts real time video.
The image of the student is captured, flipped and converted into grayscale. Within a for
loop, a rectangular box is created to outline the face and using a function called
recognizer.predict(), a numerical value on the closeness or similarity of the face is found by
comparing with the trained faces. Using the “If” and “Else” conditions, if the confidence
value is within a certain threshold, it will show the name of the face on the window, else it
will show “unknown face”. It prints the name of the student and confidence
level, and using above entry function, the name and status record is then entered into the
excel sheet and the loop is exited.
Real time surveillance can be obtained using the backend function, This function gives
produces a real time monitor which can be used for surveillance in lab facilities and classrooms
for real-time surveillance. Its implemented to show that the system works by showing a live
feed. The variables are initialized and uses Video capture function. Within an infinite loop, the
images are captured using read function, flipped and converted into grayscale. A rectangular
box is made to show that it identified the face using predict function, the confidence levels are
predicted. If the predicted confidence levels are within a certain threshold, the name of the
student is displayed. Using the escape key, the real time monitor can be exited.
CHAPTER - 5
5. 1 DESIGN PROCESS
In this research, we divide two main processes, which are the training and the testing
processes. On the training process, training sets are extracted by Eigenface. The work system
of this program uses a webcam PC to take pictures of a person's face then a face recognition
process is performed. the overall program design of the system can be seen in Figure 1 and
Figure
The process in this system starts with the mode settings that will be used.
a.System Design
User identification defines entities that are involved and interact directly with the system.
Based on the general description of the system described in the previous section, there are 2
mode who will use this application, namely:
1) Backend
that admin can enter student data and faces that will be used for the face
recognition process into the database.
2) Frontend
is a mode where the user matches the ID and face data that will be used to
proceed to the exams.
3.2 Implementation
This application is built on a desktop basis, using the Python programming language,
Emgu CV library, and SQLlite relational database management system. Emgu CV Library is
used as a tool to implement the Eigenface method that is used in the process of detection and
face recognition of students to record attendance status. And SQLlite database management
system is used as a database server to store data in the application.
CHAPTER – 6
6. 1 RESULT AND DISCUSSION
Test results on several samples. The test results in the form of the accuracy of the
face recognition program using the eigenface method for the samples tested and the level of
accuracy of the introduction and time of the testing process. In conducting testing on the
application prototype carried out on a laptop computer with the specifications of the Inter (R)
Core (TM) i5-825OU CPU @ 1.60GHz, the Windows 10 64-bit operating system, the screen
size is 1366 x 768 pixels, and 8.00 GB RAM.
The test that the researcher is doing testing uses the black box method. This method
emphasizes the suitability of the system output. The system has several stages. The first
stage, the attendance process begins with students logging into the attendance system, then
using webcam clarification on the computer. The application will detect and recognize
student faces if the face dataset in the database has the same lighting and resolution when
capturing images. When the face is recognized by the application, the application will
confirm whether the face that is recognized by the system is the face of the student
concerned. If students confirm correctly, the exam questions will come out and students take
the exam. the secondstage, the process of inputting face data on the system can be seen in the
first user selecting the input method, that is, from existing photo files or directly via
webcam, then it will be displayed on the Image Box Webcam. Thesystem will detect faces.
1 Identify
2 Identify
3 Identify
4 Not Identify
In the trials that have been carried out using the Eigen face method as well as several
things obtained, namely:
Researcher do experiment if the photo file that we entered in the system then the photo
file is also inserted into the smart phone, the system can recognize the photo file on the
smart phone. it is contradictory because it can be a hole to get into the system.
CHAPTER-7
PCA finds the data mean and principal components. It is popular as dimension reduction
procedure. The technique is usually used for maximizing variance and seizing strong .
D. Eigen face
Basically a set of eigenvector is called Eigen face which is used for face recognition of
human. Eigen face extract the main feature from image and major Eigen face is selected for face
recognition. In face recognition, Eigen face is used for represent the face image efficiently by
principal component analysis. Images can be reconstructed as like as original image from a few
number of weights and pictures . Eigen face in a face recognition system must be less or equal to
the total size of dataset. In figure 3 shows some of Eigen faces.
In this work ORL face database is used, which contains a set of faces taken between
April 1992 and April 1994 at the Olivetti Research Laboratory in Cambridge, UK. This
dataset contains 10 different images of 40 individuals. The files are in PGM format. Each
image size is 92x112, 8- bit grey levels. For masked faces we created masked face images
dataset using ORL face image and our own captured images. In this experiment, total 500
images used. 300 as a training image and 80-200 images used as test images on different
test cases. Figure 4 shows some dataset images of our experiments. All the work done
using MATLAB. Here use image processing toolbox of MATLAB 2017a version on
windows platform.
Proposed methodology is applied on our dataset images and found the resultant
images. First step of face recognition is created a training dataset which contains pre-
processed input images. After creating training set, average face is calculated for
normalized training set. For normalize the dataset, subtracting an average face from each
original face shows the normalized faces of original face image. Then calculate the Eigen
face which are s Finally test some images for recognition. Here create some masked face
images and also test them if it can be recognized.
CHAPTER – 8
8.1 FUTURE WORK
On the basis of the present research works carried out, the following are the scope of
future work for the detection and recognition. To increase the efficiency of detection and
recognition to above 90 percent. Expanding the domain of attendance system from colleges
to office and enterprises. Monitoring the arrival and departure of the employees accurately
and reliably without any proxy attendance. Using of superior algorithms like Fisher face
algorithm for facial recognition and classification to maximize the separation between classes
in the training data process.
To develop this system into a simple portable hardware system. The proposed methods
can be applied for Active student tracking, law enforcement, Information Security, Voter
verification, Access control, Surveillance and banking.
The entire system can be run on a raspberry pi along with a surveillance camera
which are the only requirements. The modularity of the system enables it to be
implemented or applied in classrooms and lab facilities for real-time surveillance or
regular attendance even under poor lighting conditions due to the Image Pre-processing
techniques which are being used in the system.
The system enables the user to store the data (attendance) automatically to the
attendance server onboard and also online which saves time and is also user friendly. Vision can
be used in Single Snap mode or Real time surveillance mode according to the convenience of
the user. Due to the efficient onboard data storage mechanism and real time surveillance Proxy
via pictures is impossible, resolving major limitation of Face recognition systems.
In comparison to reference the haar cascade classifier used in this system is way easier to
implement as it only requires a smaller dataset in contrast to the Deep Semi-NMF feature used
in the cited reference. In comparison to reference our current system is much simpler, quicker
and provides a very high detection success rate of over 70 per cent in contrast to the cited
reference.
Vision has been trained and tested over several situations of a classroom environment
and has achieved a maximum recognition accuracy of 74 per cent and above. The entire
Graphical user inter face is robust and easy to use and has been configured to be used with a
pocket-sized computer.
9.CONCLUSION
The research process that has been carried out, it can be concluded that the existence of
this application exam is carried out with an online system that aims to gain flexibility of time
and space in its implementation and clarification of face recognition methods that aim to
avoid defects. jockey / user. in the future, further application development is expected by
using updated algorithms for face recognition with a higher degree of accuracy. Trials for this
method are only carried out with the same level of lighting, not yet done at different lighting
levels and distances.
10.REFERENCES
[1] Li Cuimei, Qi Zhiliang, Jia Nan, Wu Jianhua, “Human face detection algorithm via
Haar classifier combined with three additional classifiers”,13th IEEE International
Conference on Electronics Measurement & Instruments (ICEMI), 2017
[2] Qingdong Liang, Wenting Fang, “College Student Attendance Qingdong Liang,
Wenting Fang, “College Student Attendance System based on Face Recognition”, IOP
Conf. Ser.: Mater. Sci Eng. 466 012107, 2018
[3] Kritika Shrivastava, Shweta Manda, Prof. P.S. Chavan, Prof. T.B. Patil, “Conceptual
model for proficient automated Attendance System based on Face Recognition and
Gender classification using Haar Cascade, LBPH algorithm along with LDA model”,
International Journal of Applied Engineering Research ISSN 09734562 , Volume 13,
Number 10 pp.8075-8080 © Research India Publications, 2018
[4] Felix Juefei-Xu, Marios Savvidas, “Subspace- Based Discrete Transform Encoded
Local Binary Patterns Representation for Robust Periocular Matching on NIST’s Face
Recognition Grand Challenge”, IEEE Transactions on Image Processing, Vol. 23,
No.8, 8,August 2014
[5] Ankita Digambar Patil, Dr. S. M. Mukane, Aishwarya Vijaykumar Patil, Simran Raju
Inamdar, “ Automatic Attendance system using Webcam”, et.al; International Journal
of Advance Research, Idea and Innovations in Technology”, ISSN: 2454- 132X
Impact factor: 4.295, vol. 4, Issue 2
[6] Visar Shu, Agni Dika, “Using Real time Computer Vision algorithms in Automatic
Attendance Management System”, 32nd Intenational Conference on Information
Technology Interfaces