Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Face Detection based Attendance System

using Open CV Approach
1*
V. Venkata Sai, 2Dr. G. Anuradha, 3D. Abhijith,4A. Venkatesh
1,2,3,4
VR Siddhartha Engineering College,
Andhra Pradesh, India

Abstract:- Attendance management has never been an The most widely used graphical user interface
easy task for college or hostel management. There are approach is Tkinter. It's a Python interface to the Python
many problems while taking attendance. They can take Tk GUI toolkit.
attendance in a variety of ways like the teacher taking the
attendance orally, fingerprint based attendance, sign Pandas is built on top of the Numpy library, which
based attendance system. But these are time-consuming can handle multi-dimensional arrays. It's one of the most
and difficult-to-implement processes.Also there may be well-known data manipulation packages, and it's included
chance of fraudlent attendance in girl’s hostel. So, having in almost every Python distribution. It's compatible with a
a face-detection-based attendance system helps in these variety of Python data science modules
situations.This system uses LBPH (Local Binary Pattern
Histogram) algorithm for training and detecting the Classes for manipulating dates and times may be
images. The first task is to register the students and take found in the Python Datetime package. These classes deal
their images. The second task is to train their images. The with dates, times, and time intervals.
third task is to detect their faces and mark the attendance. Our main objective is to develop a smart computer
As the face is used for taking attendance, this approach based attendance to monitor the attendance of students
removes the possibility of fraudulent attendance.This regularly and effortlessly.
system also reduces the paper work for storing the
attendance.  ORGANISATION
The following is the format of this paper: The second
Keywords:- Face Detection, Attendance Management, LBPH
section discusses literature reviews. The proposed
Algorithm, Tkinter, OpenCV.
framework is described in Section 3. And the section 4
I. INTRODUCTION describes about the Results and Analysis.

The human face plays important role in our daily lives, II. LITERATURE SURVEY
especially to identify the people. Face recognition is a type of
The whole debate in this part relies around the many
technique that takes a person's facial images and stores them
literature survey articles to which we alluded[1]. Real-time
as a unique face print to identify them. Because of its vast
computer vision methods are employed in automatic
application, face recognition technology has attracted the
attendance management systems. This study proposes a
interest of numerous researchers. As it is a contactless
novel approach to automated attendance management
procedure, this technology overtakes the problems of other
systems that incorporates computer vision techniques. The
recognition techniques such as fingerprint and iris recognition.
suggested system incorporates real-time face recognition
This systems may also detect a person from a distance.This
algorithms into an existing Learning Management System
technology is currently used at many public places to identify
(LMS), which automatically detects and registers students
criminals and also used in offices to check the regularity of
in attendance at a lecture. Combining machine learning
their employees.
algorithms with adaptive techniques of detecting face
A face recognition system is a piece of computer changes over time, the device works as a complement for
software that can be used to identify or validate a person by instructors.
comparing patterns based on theirfacial features. In this
Software that recognises faces The Lecture
system, OpenCV, Face Recognition, TKinter, Pandas,
Attendance System[2] is a system that keeps track of
Datetime libraries were used. OpenCV and Face Recognition
students' attendance in class. Face recognition is used to
are the main modules for face recognition. Using these
take attendance automatically. However, because the face
modules, the system first takes the images of the student and
detection rate is insufficient, it is impossible to accurately
save them in thedatabase, which inturn used for training
assess attendance using each face recognition result
purposes.Finally, when the system camera is turned on, the
independently. We provide a technique for accurately
system detects the faces and marks them as present.
calculating attendance based on all face recognition data
OpenCV is a computer vision, machine learning, and gathered from continuous observation in this study.
image processing toolbox that is free to use. It can recognise
products, persons, and even human handwriting in
photographs and videos.

IJISRT22NOV439 www.ijisrt.com 86
Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Attendance in the classroom is automatically recorded. III. PROPOSED SYSTEM
[3] In today's world, RFID (radio frequency identification) is
one of the most widely utilised automated identification This section describes about the architecture of
technologies. To make the most of this technology, there is a proposed system and methodology.
lot of research and development going on in this subject, and A. Architecture
many new applications and research areas will emerge in the
next years

[4] Face Recognition Using Machine Learning

Attendance Tracking Software In every firm, keeping track of
attendance is the most challenging task. In this study, we
proposed an automated attendance management system that
handles the issue of face recognition in biometric systems
under various real-time conditions such as lighting, rotation,
and scale.

[5] Face recognition utilising the LBPH algorithm is the

one we employed in our system. The LBPH stands for linear
binary pattern histogram, which compares each image's pixel
to its neighbours. Histograms are created based on the lbvp
values of those pixels. As a result, each image in the
collection is compared using this histogram. It will be
returned if it is closely matched.

[6] This research proposes a technique for calculating

each student's attendance based on continuous observations.
To assess attendance, students' whereabouts and facial
photographs are reviewed. Two cameras are put on the
classroom wall because the system's concept is so simple. The
first is to find a student's seat in the classroom, and the second
is to picture a student in the classroom using a camera. The
database will then contain attendance mapping software..

[7] This study presents a framework for dealing with

natural language processing. One is a web-based programme,
and the other is a traditional application. Both of them are
linked to a single centralised database. Face detection is
accomplished using the Haar cascade approach, while face
recognition is accomplished using the PCM method. Natural
Language Processing (NLP) is a research framework built
under SMRT-FR that allows machines to break down and Fig .1: Proposed system diagram
comprehend human language. Its purpose is to manage and
process both applications. Employees will be able to manage The proposed system clearly describes the whole
their authorizations via an interface or SMS, and the NLP process of what will be happening in the application from
programme will handle the requests. The result will be the start to the end.
handled under various circumstances. B. Methodology
This system contains various modules like:
 Registering the students and taking images.
 Training the images .
 Detecting the images and marking attendance.

After taking the images of the students, they are

converted into grayscale and stored in the dataset. The
conversion is shown in the below figure 2.

IJISRT22NOV439 www.ijisrt.com 87
Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
The feature vector of an image is obtained with the
help of LBP values that were calculated before. Now, the
image is split into smaller parts. For example, an image is
split into 64 parts and each part has a histogram constructed.
It can be seen in figure 4. Now, after all the histograms are
obtained, these are combined into one and the result is a
face vector, which in turn is a big histogram that gives us
the global description of the image.

a) Take Images
At first details of the student are entered and take
their images during the joining of the students into
the organisation or hostel. By using the Video
Capture( ) method of cv2 library we take the images
of a person standing infront of the camera until the
desired number is reached, in this system we take 60
images of each person. The captured images are now
Fig . 2: Image Processing saved in the Training image folder. When clicked on
take image the student student details are saved as
The below figure represents the LBP value. shown in figure 5.The details of students are stored
in the Student Details folder.

Fig. 3: LBP value

By using the intensity of the centre picture element as a Fig. 5: Images Saved
threshold, if the neighbour pixel has more intensity, then it is
assigned as 1 or else 0. Hence, each pixel is associated with an Figure 5 decribes about, that the images are saved for
eight digit binary number which is known as LBP value. the given Id and Name.

b) Train Images
Images are trained using the images in the Training
Image folder when clicked on train button. It is
shown in figure 6. The more the dataset is, the more
is the accuracy of detection. During training of the
images, we first get the path of all files in the folder
and then loop through all the image paths and load
the Ids and the images after creating an empty face
list and id list. This system uses LBPH algorithm for
training the images.

Fig .4: Working of LBPH Algorithm

IJISRT22NOV439 www.ijisrt.com 88
Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
On the other hand, if the image doesn’t match with
any of the images in the dataset then it displays “Unknown”
on the frame as shown in figure 8.

Fig. 6: Train Images

Fig. 8: Unknown Image
After loading, images are converted to grayscale.To
manipulate, to open and to save images we use PIL(python
image library).Then they are converted to a numpy array.
Finally, we extract id from the image, the face from the
training image sample and return the list.Images trained status
will appear on the GUI screen.

c) Detect Images
This is where the actual algorithm works to detect
images and to store the attendance. First we capture the
image of the student and convert into grayscale to
compare. Here, we use pandas library to read the csv
file which contains the details of the student. If the
image is matched with anyone in the dataset then the
name and roll number of the student are displayed on
the window as shown in figure 7.
Fig. 9: Detect without glasses

Some times there is a chance of a person wearing

accessories like glasses, masks, etc. In such cases, there
may be, a little chance of inaccurate detection. But here we
can see our code detected the image even with out glasses
as shown in figure 9.

IV. RESULTS AND ANALYSIS

This section explains the system's output and

outcomes. The final output is a GUI(Graphical User
Interface) with different options as shown in figure 10.

Fig.7: Detecting Saved image

Fig. 10: Final Output Screen

IJISRT22NOV439 www.ijisrt.com 89
Volume 7, Issue 11, November – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
After detecting the face of the student, the attendance is Pattern Histogram Method and SHA 256 Bit in the
displayed on the Gui as shown in figure 11 and also stored in Face Recognition System", International journal of
the excel sheet as shown in figure 12. research and Review, vol 7, 2020.
[6.] Face Recognition-based Lecture Attendance
System:Y. Kawaguchi, “Face Recognition-based
Lecture Attendance System,”3rd AEARU …, no.
October, 2005.
[7.] G. RoshanTharanga 1*, S. M. S. C. Samarakoon 2, T.
A.P. Karunarathne 2, K. L. P. M. Liyanage 2, M. P.
A. W. Gamage2,D. Perera. 2 (smart - fr) –“Smart
attendance using real time face recognition”.
[8.] Puja S. Prasad, Rashmi Pathak, Vinit Kumar Gunjan
and H. V. Ramana Rao, "Deep Learning Based
Fig. 11: Attendance Displayed in GUI Representation for Face Recognition", springer,
Singapore, 2020.
[9.] Ashish Choudhary, Abhishek Tripathi,Abhishek
Bajaj,Mudit Rathi and B.M Nandini Information
Science and Engineering, The National Institute of
Engineering, Automatic Attendance System Using
Face Recognition. IRJET 12/5/2013.
[10.] Anushka Waingankar1, Akash Upadhyay, Ruchi
Shah, Nevil Pooniwala, Prashant Kasambe. Face
Recognition based Attendance Management System
using Machine Learning IEEE 8/6/2012.
Fig. 12: Attendance stored in Excel sheet [11.] By Alan D. Moore , Python GUI Programming with
Tkinter Develop Responsive and Powerful GUI
V. CONCLUSION AND FUTURE WORK Applications with Tkinter , IEEE 4/11/2010.

This Face Detection Attendance System is designed to

implement the attendance activities in government schools
and colleges mainly in girl’s hostels effortlessly in an
organised manner. This can be further extended to other
private instutions, also can be used in orphanages and other
organisations.

In the future, some of the additional features can be

included like, edit & delete the student details in the GUI and
also generating excel sheet for the absentees.

REFERENCES

[1.] Dipti Kumbhar , Prof. Dr. Y. S. Angal Department

ofElectronics and Telecommunication, Smart
Attendance System using Computer Vision and
Machine Learning yogeshangal@yahoo.co.in IRJET
1/4/2017.
[2.] P. Visalakshi, Sushant Ashish Assistant
Professor ,Department of Computer Science and
Engineering SRM Institute of Science and Technology,
Chennai, Tamil Nadu, ATTENDANCE SYSTEM
USING MULTI-FACE RECOGNITION , IEEE
27/5/2016.
[3.] CH. VINOD KUMAR , DR. K. RAJA KUMAR PG
Scholar, Dept of CS& SE, Andhra University,
Vishakhapatnam, AP, India. Assistant Professor, Dept
of CS& SE, Andhra University, Vishakhapatnam, AP,
India. Face Recognition Based Student Attendance
System with OpenCV 15/12/2014.
[4.] Face Recognition based Attendance Management
System using Machine Learning. IRJET 06, June 2011.
[5.] Ellanda Purwa wijaya, Opium Salim Sitompul, Herman
Mewengkang, "Combination of the Local Binary

IJISRT22NOV439 www.ijisrt.com 90