Special Issue - 2020 International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
ENCADEMS - 2020 Conference Proceedings

Attendance Management System using Face

Recognition
1 2
Suman Kumar Jha Aditya Tyagi
Computer Science and Engineering, Computer Science and Engineering
Dr. A.P.J. Abdul Kalam Technical University(AKTU), Mangalmay institute of engineering and Technology,
Lucknow, Uttar Pradesh,India Greater Noida,India
3 4
Kundan Kumar Madhvi Sharma
Computer Science and Engineering Computer Science and Engineering
Mangalmay institute of engineering and Technology, Mangalmay institute of engineering and Technology,
Greater Noida,India Greater Noida,India

Abstract:- Facial recognition technologies have undergoes II. APPLICATIONS

large scale upgrades in performance in the last decade and A. School
such systems are now popular in field such as security and This application can use in schools because in schools
commerce. But the real difficulty is to implement an accurate traditional way of attendance marking in use where they
attendance system in real time. This is difficult to mark
attendance of large number of students present in a
call every student by roll number to check that they are
classroom. By this paper, attendance is easy by recognize present or not. This method is time consuming. During
faces of students and mark attendance. Cascade classifier and class taking attendance can take long time. This application
LBPH (Local Binary Pattern Histogram) algorithms use in can improve the standard of the school. By this system,
face recognition. This system saves time of mark attendance. students can’t mark proxy attendance. Efforts of teacher is
decrease, increase productivity and reduce human error.
I. INTRODUCTION So form these above importance of this system, it can be
Face recognition is an important application in student easily used in school.
attendance system because it saves time in marking
attendance. It is a biometric technique. This application is B. Institutions
used in where large numbers of students are present, it is This application can use in institutions because like
difficult and time consuming to take attendance on paper schools, institutions also use traditional way of attendance
one by one. So this application will help to reduce time marking where they call every student name or roll number
consumption and make easy to take attendance of every to check that they are present or not. This method is time
student just by their face recognition. In this application consuming. This application can improve the standard of
students need to register themselves by entering their roll the institute. By this system, students can’t mark proxy
number and name which is attached with their face attendance. Efforts of teacher is decrease, increase
recognition. This data is connected with data base and store productivity and reduce human error.
all students’ data. In this application LBPH (Local Binary So form these above importance of this system, it can be
Pattern Histogram) and cascade classifier used to recognize easily used in school.
faces of students.
Our brain, as a human is made to do all of this C. College
automatically and instantaneously. Computers are In college traditional way is used like school and institute.
incapable of this kind of high level generalization, so we By using this application, process of taking attendance can
need to teach or program each step of face recognition improved and save time consumption of marking
separately[4].The main motive of attendance. This application can improve the standard of
the college. By this system, students can’t mark proxy
developing this type of application is to reduce time attendance. Efforts of teacher is decrease, increase
consumption of taking attendance of students on paper by a productivity and reduce human error.
teacher during their lecture. So form these above importance of this system, it can be
Various Techniques For Marking Attendance are: easily used in school.
1. Signature based system
2. Fingerprint based system
3. Iris recognition III. RELATED WORK
4. RFID based system Types of face recognition algorithms
5. Face recognition A. PCA
Derived from Karhunen-Loeve's transformation. Given an
s-dimensional vector representation of each face in a
training set of images, Principal Component Analysis
(PCA) tends to find a t-dimensional subspace whose basis

Volume 8, Issue 10 Published by, www.ijert.org 46

Special Issue - 2020 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ENCADEMS - 2020 Conference Proceedings

vectors correspond to the maximum variance direction in example. Matching to an image involves finding model
the original image space. This new subspace is normally parameters which minimize the difference between the
lower dimensional (t<<s). If the image elements are image and a synthesized model example projected into the
considered as random variables, the PCA basis vectors are image. [10]
defined as eigenvectors of the scatter matrix. [10] I. 3-D Morphable Model
B. ICA Human face is a surface lying in the 3-D space
Independent Component Analysis (ICA) minimizes both intrinsically. Therefore the 3-D model should be better for
second-order and higher-order dependencies in the input representing faces, especially to handle facial variations,
data and attempts to find the basis along which the data such as pose, illumination etc. Blantz et al. proposed a
(when projected onto them) are - statistically independent . method based on a 3-D morphable face model that encodes
[10] shape and texture in terms of model parameters, and
C. LDA algorithm that recovers these parameters from a single
Linear Discriminant Analysis (LDA) finds the vectors in image of a face. [10]
the underlying space that best discriminate among classes. J. 3-D Face Recognition
For all samples of all classes the between-class scatter The main novelty of this approach is the ability to compare
matrix SB and the within-class scatter matrix SW are surfaces independent of natural deformations resulting
defined. The goal is to maximize SB while minimizing SW, from facial expressions. First, the range image and the
in other words, maximize the ratio det|SB|/det|SW| . This texture of the face are acquired. Next, the range image is
ratio is maximized when the column vectors of the preprocessed by removing certain parts such as hair, which
projection matrix are the eigenvectors of (SW^-1 × SB). can complicate the recognition process. Finally, a canonical
[10] form of the facial surface is computed. Such a
D. EP representation is insensitive to head orientations and facial
An eigenspace-based adaptive approach that searches for expressions, thus significantly simplifying the recognition
the best set of projection axes in order to maximize a procedure. The recognition itself is performed on the
fitness function, measuring at the same time the canonical surfaces. [10]
classification accuracy and generalization ability of the K. Bayesian Framework
system. Because the dimension of the solution space of this A probabilistic similarity measure based on Bayesian belief
problem is too big, it is solved using a specific kind of that the image intensity differences are characteristic of
genetic algorithm called Evolutionary Pursuit (EP). [10] typical variations in appearance of an individual. Two
E. EBGM classes of facial image variations are
Elastic Bunch Graph Matching (EBGM). All human faces defined: intrapersonal variations
share a similar topological structure. Faces are represented and extrapersonal variations. Similarity among faces is
as graphs, with nodes positioned at fiducial points. (exes, measured using Bayesian rule. [10]
nose...) and edges labeled with 2-D distance vectors. Each L. SVM
node contains a set of 40 complex Gabor wavelet Given a set of points belonging to two classes, a Support
coefficients at different scales and orientations (phase, Vector Machine (SVM) finds the hyperplane that separates
amplitude). They are called "jets". Recognition is based on the largest possible fraction of points of the same class on
labeled graphs. A labeled graph is a set of nodes connected the same side, while maximizing the distance from either
by edges, nodes are labeled with jets, edges are labeled class to the hyperplane. PCA is first used to extract features
with distances. [10] of face images and then discrimination functions between
F. Kernel Methods each pair of images are learned by SVMs. [10]
The face manifold in subspace need not be linear. Kernel M. HMM
methods are a generalization of linear methods. Direct non- Hidden Markov Models (HMM) are a set of statistical
linear manifold schemes are explored to learn this non- models used to characterize the statistical properties of a
linear manifold. [10] signal. HMM consists of two interrelated processes: (1) an
G. Trace Transform underlying, unobservable Markov chain with a finite
The Trace transform, a generalization of the Radon number of states, a state transition probability matrix and
transform, is a new tool for image processing which can be an initial state probability distribution and (2) a set of
used for recognizing objects under transformations, e.g. probability density functions associated with each
rotation, translation and scaling. To produce the Trace state.[10]
transform one computes a functional along tracing lines of N. Boosting & Ensemble Solutions
an image. Different Trace transforms can be produced from The idea behind Boosting is to sequentially employ a weak
an image using different trace functionals. [10] learner on a weighted version of a given training sample set
H. AAM to generalize a set of classifiers of its kind. Although any
An Active Appearance Model (AAM) is an integrated individual classifier may perform slightly better than
statistical model which combines a model of shape random guessing, the formed ensemble can provide a very
variation with a model of the appearance variations in a accurate (strong) classifier. Viola and Jones build the first
shape-normalized frame. An AAM contains a statistical real-time face detection system by using AdaBoost, which
model of the shape and gray-level appearance of the object is considered a dramatic breakthrough in the face detection
of interest which can generalize to almost any valid research. On the other hand, papers by Guo et al. are the

Volume 8, Issue 10 Published by, www.ijert.org 47

Special Issue - 2020 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ENCADEMS - 2020 Conference Proceedings

first approaches on face recogntion using the AdaBoost 2. Haar Cascade Classifier
methods.[10] Haar cascade is based on the Haar Wavelet technique to
analyze pixels in the image into squares by function. This
IV. IMPLEMENTATION uses machine learning techniques to get a high degree of
There are many types of face recognition. But we use only accuracy from what is called ”training data”. This uses
these algorithms in this application. “integral images” concepts to compute the “features”
A. Algorithm detected. Haar cascades use the Adaboost learning
algorithm which selects a small number of important
features from a large set to give an efficient result of
classifiers.
The training data used in this project is an XML file called:
Haarcascade_frontalface_default.XML
The OpenCv HaarCascade method to load the
haarcascade_ trainedfaces.txt as the classifier. The
Fig. 2 Images stored in grayscale classifier outputs a "1" if the region is likely to show the
1. LBPH algorithm: object (i.e., face), and "0" otherwise [6].
LBPH stands for Local Binary Pattern Histogram, this B. WORKING OF APPLICATION
algorithm is proposed in 2006, It is a basic algorithm that’s 1. For adding new student:
used to detect faces from front side. It is used for object as i. Enter roll number in roll number field.
well as face detection. The LBP operator helps to get local ii. Enter name in name field.
features by Local Binary Pattern acts. The local special iii. Then take image.
arrangement of the face is shortened by these LBP acts. iv. Image is capturing through the camera.
The LBP operator divides the face in the image into pixels. v. Image is processed and convert in grayscale, then stored
Every pixel is associated with 8 neighbor pixels that in database.
surroundings it. Each pixel value is then compared with the 2. For attendance of students
surrounding neighbor pixel values. The equation is for this i. Click automatic attendance button.
is: ii. Input subject name.
LBP(xc,yc) = ∑7n=0 s(in – ic)2n (i) iii. Image is capturing through the camera.
Where iv. Image is comparing to registered students.
ic - It’s the value of the center pixel v. If face is match then attendance is marking for that
(xc ,yc), - It’s value of eight surrounding pixels.[3] subject with date and time of that student.
3. For checking registered students.
FLOW DIAGRAM OF ALGORITHM i. Clicked check registered students.
ii. Enter admin id and password.
iii. Click login button.
iv. Registered students details is shown.

V. PROPOSED METHODOLOGY
A. Local Binary Patterns Histogram(LBPH)
The method proposed in this paper is marking attendance
using face recognition. Attendances is mark using a camera
and detect the faces in the image and compare the detected
faces with the student database and mark the attendance[4].
The attendance get marked in a excel file.
The application has two main parts:
1. Development of Face Recognition System.
2. Development of Attendance System.
Development of complete attendance system is achieved
using LBPH algorithm and python. Here the application
takes data like roll number, name, subject name and
provide a click to start the attendance( auto and manually).
B. Pros
1. LBPH algorithm is one of the easiest face recognition
algorithms.
2. It can represent local features in the images.
3. It is possible to get great results (mainly in a controlled
environment).
4. It is robust against monotonic gray scale transformation.
C. Cons
Fig. 2

Volume 8, Issue 10 Published by, www.ijert.org 48

Special Issue - 2020 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ENCADEMS - 2020 Conference Proceedings

1. It will require high processing power (8 GB RAM & 2

GB GC or higher)
2. If you think it will recognize person just like humans,
than leave it, It’s not possible.
3. Noisy image can reduce your accuracy so quality of
images matter.
4. It can take maximum 5-7 students at a time.

VI. CONCLUSION
In this paper, we have discussed many algorithms that are
used to make attendance. Every one have different
advantage and disadvantage.
An attendance management system is important for any
organisation. Because most of the organisations have same
traditional system which is time consuming for teachers
and students. This application is to solve this issue by
adding face recognition with attendance management
system.
The aim of this application is to make a system that is
useful to the organization such as school, college and
institute. The efficient and accurate method of attendance
in the organization that can replace the old manual
methods. This method is secure enough, reliable and
available for use. No need for specialized hardware for
installing the system in the organization. It can be
constructed using a camera and computer.

VII. REFERENCES
[1] Zhao Pei, Hang Xu, Yanning Zhang, Min Guo and Yee-Hong
Yang, “Face Recognition via Deep Learning Using Data
Augmentation Based on Orthogonal Experiments”, Article
Electronics August-2019.
[2] Ghalib Al-Muhaidhri, Javeed Hussain ,“Smart Attendance
System using Face Recognition", International Journal of
Engineering Research & Technology,
ISSN: 2278-0181, Vol. 8 Issue 12 , December-2019.
[3] Rajath S Bharadwaj, Tejus S Rao and Vinay T R,”Attendance
management using Facial Recognition”, International Journal
of innovative Technology and Exploring Engineering
ISSN:2278-3075 , Volume: 8 Issue: 6,April 2019
[4] Anushka Waingankar, Akash Upadhyay, Ruchi Shah, Nevil
Pooniwala, Prashant Kasambe ,” Face Recognition based
Attendance Management System using
Machine Learning”, International Research Journal of
Engineering and Technology ,e-ISSN: 2395-0056, p-ISSN:
2395-0072, Volume: 05 Issue: 06, June-2018.
[5] Sathyanarayana N, Ramya M R, Ruchitha C, and Shwetha H S,
“Automatic Student Attendance Management
System Using Facial Recognition”, International Journal of
Emerging Technology in Computer Science & Electronics ,
ISSN: 0976-1353 Volume 25 Issue 6 , MAY 2018.
[6] Rohit Chavan, Sankalp Sawant, Vinayak Futak,Baburao Phad
and Asha Rawat,”Attendance Management System using Face
Recognition”, International Journal for Innovation Research in
Science & Technology, volume 1,Issue11, ISSN 2349-
6010,april 2015.
[7] K.senthamil Selvi,P.Chitrakala, A. Antony Jenitha, “Face
Recognition Based Attendance Marking System ”,International
Journal of Computer Science and Mobile Computing, ISSN
2320-088X,vol 3,Issue 2,February 2014.
[8] Jomon Joseph and K.P. Zacharia, “Automatic Attendance
Management System Using Face Recognition”, International
Journal of Science and Research ,Volume 2 Issue11, ISSN:
2319-7064, November 2013.
[9] https://becominghuman.ai/face-detection-using-opencv-with-
haar-cascade-classifiers-941dbb25177
[10] https://www.face-rec.org/algorithms/

Volume 8, Issue 10 Published by, www.ijert.org 49