International Journal of Trend in Scientific Research and Development (IJTSRD)

Volume 5 Issue 1, November-December 2020 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470

Face Recognition Based Automated Student Attendance System

Ms. Pranitha Prabhakar1, Mr. Kathireshan2
1MCA Scholar, School of CS and IT, Department of MCA, 2Professor, MCA Department School of CS and IT,
1,2Jain (Deemed-to-be) University, Bengaluru, Karnataka, India

ABSTRACT How to cite this paper: Ms. Pranitha

Face recognition system is very beneficial in real-time applications, Prabhakar | Mr. Kathireshan "Face
concentrated in security control systems. Face Detection and Recognition is a Recognition Based Automated Student
vital area in the province of validation. In this project, the Open CV based face Attendance System"
recognition strategy has been proposed. This model integrates a camera that Published in
captures an input image, an algorithm - Haar Cascade Algorithm for detecting International Journal
face from an input image, identifying the face and marking the attendance in of Trend in Scientific
an excel sheet. The proposed system implements features such as detection of Research and
faces, extraction of the features, exposure of extracted features, analysis of Development (ijtsrd),
students' attendance, and monthly attendance report generation. Faces are ISSN: 2456-6470, IJTSRD38083
recognized using advanced LBP using the database that contains images of Volume-5 | Issue-1,
students and is used to identify students using the captured image. Better December 2020, pp.1346-1349, URL:
precision is accomplished in results and the system takes into account the www.ijtsrd.com/papers/ijtsrd38083.pdf
changes that occurs in the face over some time.
Copyright © 2020 by author (s) and
KEYWORDS: Face Recognition, Haarcascades, Adaboost, OpenCv, LBPH International Journal of Trend in Scientific
Research and Development Journal. This
is an Open Access article distributed
under the terms of
the Creative
Commons Attribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by/4.0)

I. INTRODUCTION
These days, biometric authentication methods begin selected image. Histogram is unique for every unique image
increasing rapidly as one of the assuring authentication and helps in determining and recognizing the physique
techniques, besides the standard authentication method. All accurately. Thus, using this technique, all images are
biometrical technologies require certain actions to be identified during the face identification phase.
entered by the user, which are the user needs to place funds
on the scanner to set the fingers or the hand geometry II. LITERATURE SURVEY
detection. The user should standstill in a fixed spot in front Traditionally attendance was taken manually which is very
of the camera for face identification purposes. The face time-consuming and had the problem of human error.
recognition method has numerous advantages when Adding on, there are numerous ambiguities towards the
compared to other biometric methods, as this method is roots of the attendance records which in fact, most of the
done passively without explicit action. This method can be attendance records are not recovered from the substantive
really useful for supervision. This project is based on the situation. The existing method uses paper or attendance
Open CV based face recognition approach has been registers for marking student’s attendance which causes a
proposed. OpenCV (Open Source Computer Vision Library) is lot of trouble. Based on various research, many resolutions
an open source computer vision and machine learning are available to resolve this concern. “Face Recognition
software library. OpenCV was built to provide a common Based Attendance Marking System” (Senthamil Selvi,
infrastructure for computer vision applications and to Chitrakala, Antony Jenitha, 2014) is based on the
accelerate the use of machine perception in the commercial identification of face identification to solve the former
products. The library has more than 2500 optimized attendance system’s issues. The system utilized a camera to
algorithms, which incorporates a complete set of both classic capture images to establish face recognition. The captured
and state-of-the-art computer vision and machine learning image was rivaled one by one with the face database to
algorithms. The experimental results show that the system examine for the students face where attendance would be
can recognize the faces captured automatically by the marked when an appropriate result was found in the
camera accurately. Thus face detection and recognition database. The foremost benefit is that the system where the
module detect faces from the image captured by the camera, attendance is recorded is actually on the server resulting in
and the image of the face is stored. The training database is high security, where a proxy cannot be given. Furthermore,
created by training the system model with the pictures of the in this proposed system, the face detection algorithm is
authorized students. The obtained images are later stored in enhanced by using skin analysis technology to enhance the
a database with appropriate labels. The features are certainty of the detection process. Although more efforts
extracted using LBPH algorithm. Haar features are basically were invested to obtain an accuracy of the face detection
used to detect the presence of particular features in the algorithm, this system required a standalone pc that needed

@ IJTSRD | Unique Paper ID – IJTSRD38083 | Volume – 5 | Issue – 1 | November-December 2020 Page 1346
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
a constant power supply making it static and hence was not III. PROPOSED SYSTEM
portable. This type of system is only proper for marking In the existing system, attendance is accomplished using the
staff’s attendance as they only need to report their attendance register. Lecturers mark the attendance of
attendance once a day, unlike students which require students manually, resulting in increased paperwork,
reporting their attendance for every class every day. It would preservation of records, and computing the attendance of
be inconvenient if the system is not portable. Thus, the each student becomes exasperating and time-consuming.
attendance management system can be established on a The intention of developing a Face recognition based
rooted design so that it can work similarly with just batteries Attendance System is to computerize the traditional way of
making it portable. practicing attendance. It assists to access the attendance
information of a specific student in a certain class. It also
The research journal “Fingerprint Based Attendance System avoids proxy of attendance. Data accuracy is affirmed, within
Using Microcontroller and LabView” (Kumar Yadav, Singh, a short span of time. The
Pujari, Mishra, 2015) recommended a solution of using a
fingerprint to mark the attendance. This system is based on FACE RECOGNITION BASED AUTOMATED STUDENT
using 2 microcontrollers to deal with the fingerprint ATTENDANCE SYSTEM demonstrates "Human Face
recognition process. Initially, the fingerprint configuration Identification using LBP and HaarCascade and openCv
will be achieved over a fingerprint sensor, then the data will Features for Real Time Attendance Monitoring. The
be transmitted to microcontroller-1. Next microcontroller-1 investigational outcomes demonstrate that the system can
will transfer the data to microcontroller-2 to check the distinguish the faces captured automatically by the camera
database. On finding a student’s match, the specifications are precisely. The results obtained from the proposed system is
sent to the PC through serial communication to be displayed. very accurate and is unambiguous resulting in accurate
This scheme is good as it expedites improvement while system. The captured images are then detected and
sustaining design flexibility and simplifies testing. But again, compared with the images present in the database and the
this system is connected to a PC which makes it unportable. attendance is then marked.
Apart from that, the database information cannot be
accessed, resulting in a situation where parents who are IV. SYSTEM DESIGN
curious about knowing their child’s attendance cannot easily ARCHITECTURE OF THE SYSTEM:
access the data. Therefore, to provide accessibility to the
legitimately concerned party, the data can be uploaded to a
web server for easy access. The authentication for the
appropriate access can be implemented through a login
screen. According to the research journal “Attendance
System Using NFC Technology with Rooted Camera on
Mobile Device” (Bhise, Khichi, Korde, Lokare, 2015). The
attendance system was implemented using Near Field
Communication (NFC) technology and mobile application.
According to the research paper, every student was given an
NFC tag that contained a unique ID during their enrolment
into the college. Attendance of each class was then taken by
touching or placing these tags on the lecturer's phone. The
rooted camera on the phone then captured the student’s face
and the data was directed to the college server for validation
and verification. The benefit of this method is that the NFC is FACE DETECTION: The principal utility of this step is to
uncomplicated to use, and the speed of connection accomplish whether the human faces appear in a given
establishment is very high. However, the system couldn’t image, and what is the position of these faces. The
automatically spot any violations when the NFC tag is not predictable outputs of this step are squares which contain
carried or tagged by the authenticated owner. The NFC each face in the input image. Face orientation is performed
reader was an inconvenience to the lecturer. Most of the to vindicate the scales and alignment of these squares.
lecturers did not like to prefer their smartphones to be used FEATURE EXTRACTION: Following the face detection step
in this way due to privacy concerns. Real Time Human Face the extraction of human face areas from images that are
detection and Tracking was proposed by J. Chatrath, P. done. FACE RECOGNITION: The last step after the
Gupta, P. Ahuja, A. Goel. Their paper outlines the practice of illustration of faces is to identify them. For automatic
real time profile detection and acknowledgement by acknowledgement we need to build a face database. Various
modifying Viola-Jones algorithm. Outcomes accomplished by images are taken for each individual and their features are
the developed algorithm displayed that up to 50 faces could mined and stored in the database. Then when an input image
be distinguished and tracked by systems using the modified is served, the face detection and feature extraction is
algorithm. Processing of data and time consumed is performed and its feature to each face class is compared and
moderately less in such systems. Implementation of stored in the database.
Attendance Management System was proposed by G.
Lakshmi Priya and M. Pandimadevi. Systems proposed with ALGORITHM: Local Binary Patterns Histograms - This
respect to this idea was to capture an image using a web method prerequisites the gray scale pictures for dealing with
camera at deviating occurrences. A precision of 68% was the training. LBPH uses the following parameters: Radius,
observed in such systems. Neighbours: The number of sample points neighbouring the
central pixel which is commonly 8, Grid X: The number of
cells along the horizontal direction is denoted as Grid X, Grid

@ IJTSRD | Unique Paper ID – IJTSRD38083 | Volume – 5 | Issue – 1 | November-December 2020 Page 1347
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Y: The number of cells along the vertical direction is denoted
as Grid Y.

ALGORITHM TRAINING: For the training resolution of the

dataset of facial images of the people to be predictable
alongside with the unique ID, is compulsory so that the
existing approach will utilize the delivered information for
recognizing an input image and providing the output. Same
images need the same ID.

EXTRACTION OF HISTOGRAM: The image obtained in the

preceding step practises the Grid X and Grid Y factors and
the image is split into numerous grids. Based on the image,
the histogram can be mined as below: The image is in grey
scale and each histogram will contain only 256 positions (0-
255) which indicates the presences of each pixel
concentration. After this each histogram is formed and a new Simple GUI is made via the python Tkinter library which
and bigger histogram is done. Let us presume that there are permits the user to use the page without any backend
8×8 grids, then there will be 16.384 locations in total in the familiarity. Tkinter is the standard GUI library for Python.
final histogram. Ultimately the histogram indicates the
features of the actual image. Haar Cascade Object Detection
Face & Eye OpenCV: Haar cascade classifier is based on the
Viola-Jones detection algorithm which is trained in giving
some input faces and non-faces and training a classifier that
recognises a face. So we need to gauge a huge set of features
for every 24*24 PX. So to evade this AdaBoost is used.
AdaBoost is used to eliminate redundant features and
choose only appropriate features. These features are also
termed as weak classifiers. After these features are found a
weighted mixture of all these features is used in assessing
and determining if any given frame has a face or not.

Strong classifier = linear sum of weak classifiers

F( x ) = ∑ ( αᵢ * fᵢ( x ) ) [here αᵢ are matching weights to each

weak classifier fᵢ(x).]

A Haar Cascade is based on “Haar Wavelets” which defines After recognizing all faces of the students, the names of
as: An arrangement of rescaled “square-shaped” functions persons are updated into an excel sheet.
which composed form a wavelet family or basis. It is built on
the Haar Wavelet technique to analyze pixels in the image Whose faces are not recognized due to various external
into squares by function. This uses machine learning factors are stored as “Unknown images” in the database.
practises to get a high degree of accurateness from which it
is called “training data”. V. IMPLEMENTATION
The face recognition attendance system will be installed on
the host device and all the permissions will be granted
during the installation. Once the installation is over the
administrator can access the system and authorize the user
and dictate the image and record the image and mark
attendance and the data stored is database using a
specialized database. One can give their attendance by
merely facing the camera. When we run train.py a window is
opened and requests for entering Id and Name. After
registering name and id then click the "Take Images" button.
By clicking the "Take Images", the web camera of the host
computer is opened and starts clicking an image sample of
the person. This Id and Name is stored in the folder Student
Details and the file name is StudentDetails.csv. After
conclusion, it notifies that images are saved. After taking an
image sample, click the "Train Image" button. Now it takes a
few seconds to train the machine for the images that are
Feature Extraction Haar Cascades use machine learning taken by clicking the "Take Image" button and creates a
methods in which a function is trained from a lot of positive Trainner.yml file and stores in the Training Image Label
and negative images. This process in the algorithm is feature folder. Now all initial setups are done. If the face is
extraction. recognized by the system then the Id and Name of the person

@ IJTSRD | Unique Paper ID – IJTSRD38083 | Volume – 5 | Issue – 1 | November-December 2020 Page 1348
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
are shown on the Image. Press Q (or q) to quit this window. VII. FUTURE ENHANCEMENTS
After exiting the window, the attendance of the student will Further expansions can be performed, to attain the real time
be stored in the Attendance folder as a CSV file with name, id, detection of the specific student in the monitoring premises.
date, and time. After identifying all faces of the students, the Alternatively to taking images, we can also work with
names of individuals are updated into an excel sheet. Whose recorded videos. But some period is managed to record the
faces are not recognized due to various external factors are images, because if a continuous recording is done, then load
stored as “Unknown images” in the database. on database increases. The future output is to develop the
recognition rate of algorithms when there are accidental
VI. FLOWCHART changes in a person like a tonsuring head, using a cap, having
a beard, etc. The system manifested only recognizes face up
to limited angle variations which can be developed further.
Gait recognition can be combined with face recognition
systems to achieve better performance of the system.

VIII. CONCLUSION
This method helps to evade the fail-proof attendance system
and works as a replacement for all present systems i.e. other
bio-metric practices. It's time-saving and consumes low
manpower with respect to taking attendance manually.
Automated Attendance Systems based on face recognition
techniques thus proves to be time-saving and secure. This
system can also be used to identify an unknown person, thus
being able to recognize if the student actually belongs to the
respective class or school. Hence, proving that it's an easy
method to take attendance and store as well.

IX. BIBLIOGRAPHY
[1] https://docs.conda.io/en/latest/
[2] https://opencv.org/about/
[3] https://docs.opencv.org/3.4/db/d28/tutorial_cascad
e_classifier.html
[4] https://towardsdatascience.com/the-intuition-
behind-facial-detection-the-viola-jones-algorithm-
29d9106b6999
[5] https://www.ijedr.org/papers/IJEDR1903093.pdf
[6] Tiwari PritiAnilkumar, KalyaniJha, Karishma P Uchil,
Naveen H., "Haar Features Based Face Detection and
Recognition for Advanced Classroom And Corporate
Attendance", IJIRCCE,Vol. 3, Issue 5, May 2015.
[7] Venkata Kalyan Polamarasetty1, Muralidhar Reddy
Reddem2, Dheeraj Ravi3, Mahith Sai Madala4”
Attendance System based on Face Recognition”
International Research Journal of Engineering and
Technology (IRJET) e-ISSN: 2395-0056 Volume: 05
Issue: 04 | Apr-2018 www.irjet.net p-ISSN: 2395-
0072

@ IJTSRD | Unique Paper ID – IJTSRD38083 | Volume – 5 | Issue – 1 | November-December 2020 Page 1349