International Journal of Recent Technology and Engineering (IJRTE)

ISSN: 2277-3878, Volume-9 Issue-1, May 2020

Automatic Attendance System Using Face

Recognition Technique
Mayur Surve, Priya Joshi, Sujata Jamadar, Minakshi Vharkate

Eigenfaces, LBPH, fisherfaces, SIFT these are some face

Abstract: Attendance system is very important in schools and recognition techniques. For face detection AdaBoost, SVM
colleges’ Manual attendance system has many difficulties like it based, Viola-Jones face detector these algorithms are used.
may less accurate and critical to maintain. So, attendance system In our implementation we are using Haar cascade algorithm
using face recognition technique increase the accuracy and also it
and AdaBoost classifier for face detection and recognition. It
required less time than other methods. There are many existing
system for attendance such as face recognition using IoT, PIR is divided into four steps:
sensors and so on. For face recognition, hardware devices also Haar feature: It collects the features and expression of faces. It
helpful. But challenge is that to maintain all the sensors properly takes all features by analyzing all images which it captures.
without get damage. After studying all method and techniques we Internal Image: After collecting all features it creates internal
are trying to implement a system with Haar Cascade Algorithm image using line feature, edge feature, four rectangle feature,
which has highest accuracy among all. It is able to capture the etc. these are some pixel calculations.
images from 50-70cm. We are creating graphical user interface Ada boost classifier: In this step all features from faces and
which capture the images, create the dataset and train the dataset uniqueness are combining together.
on single click. After recognizing the face it will display name of
Cascading classifier: In this step it will classify the faces and
student and roll number. That information stored in attendance
sheet automatically with time and date. non faces successfully. It is for testing purpose.

Keywords: face detection, face recognition, haar features, II. LITERATURE REVIEW
histogram of oriented gradient
Face and Bio-Metric Based Attendance and Security
System using RFID and Arduino
I. INTRODUCTION
In this paper they had proposed an employee attendance &
As we have seen in many schools and colleges that faculty security system using face and Biometric integrated with
members are facing the problem of proxy attendance, Smart RFID cards using Arduino. The project was about
maintaining all hand written document of student attendance implementation of RFID based attendance system in
of each batch/class every day. It is very difficult task for them. integration of face recognition of the particular employee with
To reduce their work many techniques are come into picture. his unique employee number. With integration of finger print
RFID system: In this system student has to carry RFID card authentication (Bio-metric) into the system, security has been
for attendance purpose every day. enhanced. Audio welcome message on the valid Employee
That card consists of chip by swiping this into card reader the attendance registration had introduced & for unauthorized
system will read all details of student and marks the entry Audio alert had introduced along with Sound Alarm[2].
attendance.
Biometric system: It will scan the unique part of body such as Attendance System based on face recognition
fingerprints. At initial we have to store fingerprints of each This paper had focused to improve attendance system in
candidate in database. During attendance present fingerprint school and colleges. Because there were many disadvantages
and stored fingerprint get checked. in taking manual attendance like cost, fake attendance and it
Bluetooth system: It is used for removing proxy attendance may be not accurate always. So biometric and face
over limited population. But it allows only eight connections recognition techniques come into picture. Traditional face
at a time. recognition methods were not more accurate. This paper had
This above systems totally based on hardware device. In case represented the idea that to save the time as well as enhance
of software system many face recognition techniques are used the security Attendance System using face recognition
for attendance. technique was very useful. In this idea images were capturing
through camera or CCTV camera in school/college for
attendance purpose. After that it first detected the face using
features of eyes, nose, mouth, hairs and also different pose of
face in image. There were different methods for face detection
Revised Manuscript Received on May 16, 2020. like Ada-Boast, SQMT, LBP and SNOW classifier method.
Mayur Surve, Student, Department of Mechanical Engineering in MIT Once face was detecting apply face recognition method
Academy of Engineering, Pune, India. Histogram of Oriented Gradient features (HOG), Haar
Priya Joshi, Student, Department of Mechanical Engineering in MIT
Academy of Engineering, Pune, India. Cascade features.
Sujata Jamadar, Student, Department of Mechanical Engineering in
MIT Academy of Engineering, Pune, India.
Mrs. Minakshi N. Vharkate, Sr. Assistant Professor, Department of
Computer Science and Engineering, School of Computer Engineering and
Technology, MIT AOE, Alandi (D), Pune, India .

Published By:
Retrieval Number: A2644059120/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A2644.059120 2134 & Sciences Publication
 Classroom Attendance System using Face recognition

Then it will compare the captured images with database. If it Attendance Management Using Facial Recognition
will match attendance is marked. If not then that image will Approaches for attendance till date
store in unknown person database. But this method used only  Biometric scan: Students’ finger print scan and stored
for small database. in database. To mark attendance student have to use
The result of detection and recognition was come out using biometric system before every class.
MATLAB [3]. Below block diagram shows the process and  ID card swipe system: RFID chip fitted in ID card of
working of attendance system based on face recognition by students which are scanned by system and read that
using method :
chip. That card must scan in every class to mark
 Histogram of Oriented Gradient features (HOG),
attendance. The present and absent student list will
 Haar Cascade features.
displayed to faculty[1].

Store basic Capture the Existing Algorithms

information images of student  Eigen faces: This algorithm used to extract necessary
of student in and train features of faces/images and encode it. For obtaining all
database necessary factors of faces it takes number of images &
create the set. Using that set variance and co-variance
matrix are calculated. Each image represent eigen vector.
HOG features Applying b-box  Line Edge Map: This algorithm used to extract the
are used for method for face features of eyes, nose and mouth as they contain different
face detection more features. Using those features line mapping is done
recognition on faces. Here color images are converted to gray images
as it uses gray scale method.
 Histogram of oriented gradient: This technique used for
both objects and faces. All images converted to grayscale
Compare the Mark
and each pixel assigned to integer. Main aim is to find the
images with present/absent
darker region of face.
database according to
database Student Attendance Marking Using Face Recognition in
Internet of Things
Fig 1.1 Block Diagram The basic idea was same behind this solution that was
improving the attendance system by using face recognition
Automated Attendance Mgmt System Using Face technique. This will reduce proxy attendance and increase the
Recognition accuracy.
Introduction Hardware Requirements:
The basic concept behind the face recognition is image  Microcontroller: Here Arduino Uno is used as
processing. There are two types of image processing: microcontroller which is platform for implementing
Analog processing: It is technique which used for hard copies the sensors.
such as photographs and printouts.  PIR Sensor: Passive Infrared Sensor is used for
Digital processing: It manipulates the content of digital measure the radiation from object and also motion of
images by PC. Basically attendance is of two types: object.
 Manual Attendance System
 Automated Attendance System In this proposed system face recognition had done by using
Arduino and sensors. Overall concept in this paper divided
There were many difficulties in manual attendance like into several steps.
maintain all papers, mark present/absent by pen everyday so it  At initial step student filled the registration form along
takes more effort & it more time consuming. But this all with all the details, which stored in college database.
limitation removed by automated attendance system. The Image/picture of student also stored in database. This
main objective of paper was to gain the images or videos of step required only once.
student face, their position and attendance. There were some  Then camera was going to set at entrance of class along
systems for attendance purpose but they had few
with sensor and microcontroller. Here PIR sensor used
limitations[4]:
for measure the radiation from object. It also detected
 Biometric based system: It scans the unique part of
motion of objects.
body like finger prints and mark the attendance. But it
 When student enter into the class first PIR sensor had
also time consuming.
measured the radiation of object and motion. If
 Bluetooth System: This system is not scalable and
radiation belongs to human category then camera
requires 8 connections at a time.
become activates and captures the pictures.
 RFID System: In such system each person have to use
 After clicking the images of student it compares that
RFID card for attendance for swiping. But it has
image with database if
difficulty when RFID card get lost.
image matched then it
updates the database

Published By:
Retrieval Number: A2644059120/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A2644.059120 2135 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-9 Issue-1, May 2020

with present. Otherwise marked as absent in database. That information stored in file. After it creates the dataset of
Every updation in database about particular student student face and train it. Then we are applying Haar feature &
sent to their parents through mail or sms every day. Ada boost classifier which are face recognition technique. It
 This system also worked for faculty in same manner. captures the live images from video stream & store in file
For faculty there was another database had created. while storing it compare the images in file whether they are
Using this system we could also maintain faculty matching or not with registered student. If yes then it will
attendance[5]. mark present with time and date otherwise image stored in
unknown file.
III. WORKING SYSTEM
Our solution is implemented by Haar features and AdaBoost
classifier. Here we are creating graphical user interface (GUI)
for storing name and roll number of each student in file. While
Information gathering of students it also generates dataset of
student faces and stored in folder. After completing that,
trained images also stored in folder which required for face
Recognition. These are some essential and important steps in
our implementation. In real time scenario when we place the
camera near to classroom door which capture live data
Continuously. It captures images from live streaming of that
camera. Fig 1.3 System’s Dashboard
Captured images are comparing with images which are stored
in folder at time of registration. If image match appropriately Here is our system look like, Fig1.3 shows our system
it shows student name and roll number on face which has been dashboard in which we have to enter ID and NAME of
detected or recognized. student. After entering the ID and NAME of student we have
This information automatically stored in attendance sheet to click on Icon named as Take Images. Then it will take
along with date and time. If images are not matching then, that image of student as shown in Fig1.4.
images stored in unknown folder. It is very helpful attendance
system for faculty member because it stored the attendance of
student with date and time. So, faculty members easily
identify which students attend the lecture on what time. Given
proposed system also useful for laboratory attendance,
government office attendance, institutional/organizational
attendance and library attendance.

User Detail of student

interface/GUI stored in file

Fig 1.4 System Taking Photos

Apply Haar Creating dataset By using haar features and adaboost we have train and classify
feature & & train images images. There is another icon named as Train Image. So,
Adaboost After taking image of student we have to click on Train Image
which will directly get assign to ID and NAME which we
Live image Image matching have entered as shown in Fig1.4.
capturing with stored file

(if no)Stored in
unknown If image
DB/folder matched

(if yes)Stored in
attendance sheet Fig 1.5 Face has recognized
with info
Fig 1.2 Block Diagram of System
Above block diagram explains the flow of our working
system. First student register all information through GUI.

Published By:
Retrieval Number: A2644059120/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A2644.059120 2136 & Sciences Publication
 Classroom Attendance System using Face recognition

Further is that we have to track image so we have to click on automatically stored in attendance sheet along with id, name,
Icon named as Track Image. After tracking particular image it date and time.
will show the name and Id of student on his/her detected face
as shown in Fig 1.5. then the information will stored in IV. RESULT
database. After completing all this steps we have to click on
All information of student present in image stored in
Icon named as Quit so the information of student will stored in
attendance sheet. So this system handles multiple faces at a
attendance as shown in Fig1.6 ,and it will generate an excel
time properly. Attendance sheet will update as shown in
file which have been stored with students information who are
image.
present in class. This how the system works.

Fig 1.6 student’s attendance has marked

Let us understand the system by considering the different

scenario to make it simple:
Fig. 1.8 Multiple students attendances has marked
1. If only one student enters the classroom then how the
Above Fig. shows Information of student which have been
system works:
stored into excel file after recognizing their faces by Fig.1.7
and Fig. 1.6. It stores all information of student either it is
If one student entered in classroom at time then the images of
single face or multiple faces. System works properly in all
that particular student will get capture by camera and then that
conditions. When student enter into class (single/multiple) if
Image will trained to recognize the face of the students on the
they recognize their id and name stored in attendance sheet
basis of the stored images in the dataset.
along with current date and time.
On the basis of recognition of the face of students as shown in
Fig1.8 the information related to that particular image will
V. ADVANTAGES
generate an excel file and stores students name, Roll no, Date
and time. Then system will mark the students as present. 1. It can handle large database and store large number of
One the data has been stored into the system the aim of the images for training
system is to give attendance to the present students. 2. Accuracy of this algorithm which have been used is more
than other algorithms
2. Consider the second scenario in which multiple student 3. It is able to capture images from 60-80cm distance very
enter into classroom simultaneously. In that case also it works accurately
successfully. 4. Suppose in case a student enters multiple roll number by
mistake at the time of face recognition it shows all roll number
in output
5. Network connectivity is not required so there will be no
network related problems
6. Time consumption of creating dataset and train the image is
very less.
7. Human and machine direct interface is less and which
reduces many errors and increases accuracy up to great extent.
8. System is simple to operate and also simple to operate with
good accuracy while having the attendance of students.
9. For multiple faces also system work properly & attendance
update successfully.
10. Image capturing speed also good which takes images
Fig.1.7 Multiple student face recognized
without get struck.
It can detect or recognize multiple face at time while entering
in the class, at shows name of student as shown in above Fig.
After recognition, when we click on exit button it will show
attendance notification in attendance box. That attendance

Published By:
Retrieval Number: A2644059120/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A2644.059120 2137 & Sciences Publication
 International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-9 Issue-1, May 2020

VI. CONCLUSION Sujata Jamadar: A student pursuing BTech in

Computer Engineering in MIT Academy of
Basically this system work for improving attendance system Engineering, Pune from 2017. Published a
in every domain like schools, colleges, organizations, research paper on underground optical fiber
communication safety an idea selected at national
institutions and companies. Capturing live images from level SIH (Smart India Hackathon) by ministry of
camera and applying different techniques of face detection communication. Certification in Project
and face recognition which will reduce manual or traditional Management for Managers. Got selected in SIH
work. In our solution, by creating interface we generate the software a national level competition in year 2019. Worked on various
projects like parking slot detector , website for hotel management system ,
dataset. We trained the images using Haar Cascade and automatic garbage collector ,and having interest in cyber security things.
AdaBoost classifier. After completing training it will
successfully detect and recognize faces and non faces. When Mrs. Minakshi N. Vharkate completed BE and
stored images and compared images matched then attendance ME and started her PhD in Computer Science and
Engineering. She is working as Sr. Assistant
sheet get updated automatically with time and date. As it
Professor in School of Computer Engineering and
stored the entering time of every student it becomes easy for Technology at MIT AOE, Alandi(D),Pune, MH,
faculty member to keep track on time of student. India. Her research areas are Machine Learning,
neural network, deep learning, Remote Sensing
image Processing.
REFFERENCES
1. Rajath S Bharadwaj, Teju S Rao, Vinay T R, “Attendance
Management Using Facial Recognition”, IJITEE, Vol: 8, Issue: 6,
April 2019
2. Narendar Singh, M Kusuma Sri, K. Mounika, “IoT Based Automated
Attendance with Face Recognition System”, Vol: 8, Issue:6S4, April
2019
3. [Link] Kalyan, Muralidhar Reddy, Dheeraj Ravi, Mahitha Sai,
“Attendance system based on face recognition”, IRJET, Vol: 5, Issue:
4, April-2018
4. Ali Akbar, Chowdhary Obaid, Choudhary Yasir, “Automated
Attendance Management System Using Face Recognition” IJARCET,
Vol: 6, Issue: 8, August 2017
5. Reetha. S, Dr. P. Visu, “Student Attendance Marking Using Face
Recognition in Internet of Things”, IJCST, Vol: 5, Issue: 3, May-June
2017
6. [Link] Chavan, Baburao Phad, Sankalp Sawant, Vinayak Futak, Asha
Rawat, “Attendance Management Using Face Recognition” , IJIRST,
Vol: 1, Issue: 11, April 2015

AUTHORS PROFILE
Mayur Surve: A student pursuing BTech in
Mechanical Engineering in MIT Academy of
Engineering, Pune from 2017. Published a
research paper on underground optical fiber
communication safety an idea selected at national
level SIH(Smart India Hackathon) by ministry of
communication. Currently working on Machine
Vision System as a mini project. Successfully
completed the pulse delay generator specially used in the high speed camera.
Working on " Industry 4.0: current scenario in indian manufacturing
industries". Has a working experience of 2 Months as an intern at Mercedes
Benz India Limited, Chakan,Pune. only manufacturing plant in india.
Certification in the course Project Management for managers, Cooperation
in Autonomous Vehicles, Thermodynamics, Sensors for Autonomous
Vehicles by IEEE. Mainly interested in sales and marketing, human resource
and mechanical design area, AI systems and Big Data.

Priya Joshi: A student pursuing BTech in Computer

Engineering in MIT Academy of Engineering, Pune
from 2017. Worked for social services as a teacher.
Part of SAKSHAM NGO. Published a research paper
on underground optical fiber communication safety
an idea selected at national level SIH (Smart
India Hackathon) by ministry of communication.
Currently working on the Machine vision system as a
mini Project. Got selected in the SIH (Smart India Hackathon) for the Drug
delivery project. Successfully completed the project like SMART DUSTBIN
based on Arduino and message delivery Certification in the course Project
Management for managers, Cooperation in Autonomous Vehicles,
Thermodynamics, Sensors for Autonomous Vehicles by IEEE. Mainly
interested in sales and marketing, human resource and computer science,
cyber security, AI systems and Big Data.

Published By:
Retrieval Number: A2644059120/2020©BEIESP Blue Eyes Intelligence Engineering
DOI:10.35940/ijrte.A2644.059120 2138 & Sciences Publication