© 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

Face-Recognition Attendance Management System

Mukesh Rajput1 Thakurendra singh2

M.Tech Student of ECE Deptt, Assistant Professor of ECE Deptt,

R.B.S. E.T.C.Bichpuri Agra, R.B.S.E.T.C.Bichpuri Agra,
India India

ABSTRACT

Higher education institutions are currently concerned about students' attendance trends. Even in an epidemic situation,
poor attendance is a serious issue in schools and universities. The two primary conventional methods for keeping track of
attendance are calling out the roll call or having pupils sign a piece of paper. Both of these demanded more time and effort.
Consequently, a computer-based student attendance management system is required, allowing the teachers to automatically
keep track of attendance. Using "TKINTER" and "PYTHON," we created an automated attendance system for this project.
Our plans to construct an "Automated Attendance System Based on Facial Recognition" have been predicted. The
application contains face recognition, which saves time. Also, because it is entirely software-based and uses less paper, it can
be deemed environmentally beneficial. Because this system uses facial recognition as a biometric for verification, it also
completely eliminates the possibility of false attendance. So, this technique can be used in a profession where attendance is
crucial. The suggested system is built on the TKINTER platform and supported by a PYTHON script and a SQL database.
The algorithm employed by the system is based on image comparison using the encoded values of the face from the database
image and the image captured by the system in real-time. An excel sheet is the system's output format.

Keywords: Attendance system, &Machine Learning, My SQL

INTRODUCTION
The Attendance System with Face Recognition is a modern alternative to the outdated system of recording attendance. The
suggested system is a python-based, tkinter-based system that uses a MySQL database for support. A single faculty system of a
certain institute may use this method. It is suggested that this system be built on biometrics, specifically Face Authentication. This
approach totally eliminates the possibility of faking attendance, which is a concern with traditional methods of attendance because
biometrics are used.
Each institution follows the attendance management process, which is crucial for monitoring students' development. Every institute
takes a distinct tack on this. While some of their schools still use the archaic paper- or file-based systems, others have implemented
automated attendance strategies that make use of biometric technology. By comparing patterns based on a person's facial features,
a face recognition system is a component of digital software that can recognise or confirm a person. At the time of attendance, when
the system camera turns on, the system will detect the faces that were present in the frame using HOG, or (Histogram of Oriented
Gradients), which is a popular library for face detection. In this system, the faces were detected by using Open CV & Face
Recognition libraries, which are one of the popular libraries for face detection. If the image in the frame was tilted, the Face
Landmark Estimation method will be applied and the face will be adjusted to be as near to perfectly centred as possible. The system
will then encrypt every image that was stored in the database, along with any features that were recognised in the image. For the
Deep Conversional Neural Network algorithm to execute the encoding, 128 measurements were generated for each face. The
dimensions of the features in the image that was previously stored in the database will then be compared to the dimensions of the
faces in the frame that were detected.
In the end, the system will use a basic linear SVM technique to locate the person in the database of known individuals (i.e., the one
that was captured at the project's beginning) whose measurements are closest to the image that the camera picked up. The system
will generate a name, date, time, and present mark after determining a perfect match, and it will store this information in a CSV
file. They were subsequently uploaded to the database and can be opened by users using Microsoft Excel.

LITERATURE REVIEW
Automated attendance management systems that use real-time computer vision algorithms this study suggests a new approach for
automatic attendance management systems that are enhanced by computer vision algorithms. The suggested system uses real-time
face recognition algorithms connected to an existing learning management system to automatically identify and register students
who are present at a lecture. (LMS). By combining machine learning algorithms with adaptive techniques for observing facial
changes over an extended period of time, the technology serves as an additional teaching tool for teachers.1

1
Dipti Kumbhar , &Prof. Dr. Y. S. Angal Department of Electronics and1 Telecommunication, Smart
Attendance 1System using Computer &Vision and Machine Learning yogeshangal@yahoo.co.in IRJET
1/4/20170]

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 492
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

The Lecture Attendance System, which tracks pupil attendance during lectures, is based on facial recognition technology. The
technology automatically records attendance using face recognition. It is unfortunately difficult to estimate the attendance correctly
using each face recognition result individually due to the low face identification rate. Using all the facial recognition data from
ongoing observation, we present a method for precisely estimating attendance in this paper.2
Automatic monitoring of students attendance in classrooms one of the automated identification techniques that is becoming more
frequently used today is RFID. Radio frequency identity is commonly referred to as RFID. There is a lot of research and
development being done in this area to fully utilize this technology. Numerous new uses and study fields will keep emerging in the
years to come.
Facial Recognition Attendance Management System Based on Machine learning in every business, keeping track of attendance is
the most difficult task. In this research, we addressed the issue of face recognition in biometric systems exposed to a variety of real-
time settings, such as lighting, rotation, and scaling. A computerized approach to handle attendance has been suggested.

METHODOLOGY
Every institute follows the attendance management process, which is crucial for monitoring students' development. Every institute
takes a different tack on this. While some of their institutions still use the archaic paper- or file-based methods, others have
implemented automated attendance strategies that make use of biometric technologies.
A piece of digital software called a face recognition system may recognize or confirm a person by comparing patterns based on
their facial features. HOG, or (Histogram of Oriented Gradients), a well-known library for face detection, is used by the system to
identify the faces that were visible in the frame at the time of attendance when the system camera comes on. One of the most used
libraries for face recognition, Open CV & Face Recognition, was used in this system to detect faces. If there were any tilted images
in the frame, the Face Landmark Estimation technique would then be applied, and the face would be altered to be as close as possible
to perfectly canter .The system will then encrypt every image that was stored in the database, along with any faces that were
recognized in the image. For the Deep Conversional Neural Network technique to execute the encoding, 128 measurements were
obtained for each face. The dimensions of the faces in the image that was previously saved in the database will then be compared
to the dimensions of the faces in the frame that were detected.
The system will ultimately employ a fundamental linear SVM technique to locate the subject in the database of known subjects
(i.e., the one who was photographed at the start of the project) whose measurements are most similar to the image that the camera
picked up 1 the system will generate name date time and present mark after recognizing a perfect match and it will store this
information in CSV file. They were subsequently uploaded to the database and can= be opened &by user with the help of Microsoft
excel

SUGGESTIVE SYSTEM
Image acquisition:
An HD camera that is installed in* the lab or classroom is used to (acquire images. The system receives this image as input). The
title and author 1information must be centred and in a single@column.
Dataset Creation:
A user set of data is produced preliminary to the recognition process. To train this system, a dataset of the complete class will be
created, containing each student's name, roll number, department, and images of them in various postures. Deep learning is used to
compute 128-d facial features for each face after student data and photographs are registered in our system to create a dataset. These
1features are then stored in student face data files so that the face may be recalled during recognition procedures. This procedure is
applied to each image that is taken during registration.
Face Detection and Extraction:
Face detection is important because it allows 1the system to recognize the human faces in an image that was captured by a camera.
It is now possible to locate faces within an image and recognise them using a1 variety of image processing methods. For the purpose
of locating individuals in the given image, we used the HOG technique.
Face Positioning:
There are 68 different areas on the human complexion. Alternatively known as 68 facial landmarks. The main goals of this phase
are to locate the image and identify facial landmarks. Python scripts are utilized to automatically locate the face and position it as
precisely as possible with no distorting the image.
Face encoding:
After faces have been detected in the source picture, the process of extracting the particular identifiable facial characteristic for each
image is known as face encoding. The 128 fundamental facial points, which are highly accurate, are in essence retrieved for each
picture used as input whenever we acquire a 1accurate location of the face and are then saved in data files for face recognition.
Face matching:
The Face comparison is the last step in the face recognition procedure. One of the best methods for learning, deep metric learning,
was chosen because it creates vectors of features with actual value and is incredibly accurate. To verify each face, the proposed
system produces a 128-d embedding (ratification). The internally compare faces method is used to calculate the Euclidean distance
between each face in the dataset and each face in an image. If the present image matches the existing dataset by 60% or more,
attendance marking is going to begin.

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 493
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

Attendance Marking:
Python generates the roll numbers of present students and returns them after a face is recognized from an image preserved in a SQL
database. The system builds an attendance table from the data it gets that lists each entry's name, roll number, date, day, and time
as well as the topic id. The data is subsequently transmitted to Python, which immediately converts the dataset into a CSV file.
Letter staff are able to access the file and open it up in Excel to update and change the data in the spreadsheet.

Figure 1. ER-DIAGRAM,

The suggested system's Entity-Relation diagram is depicted in Figure 1.

Activity Diagram in Figure 2

The proposed system's activity diagram is shown in Figure 2.

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 494
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

USE-CASE DIAGRAM, FIG. 3

The suggested system's user interaction diagram is shown in Figure 3.

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 495
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

FIG. 4 (LVL 0 & LVL 1)

DATA FLOW DIAGRAM The data flow diagram on levels 0 and 1 is shown in Figure 4.

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 496
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 497
 © 2023 IJRTI | Volume 8, Issue 4 | ISSN: 2456-3315

FIGURES 4, 5, 6, AND 7:* OUTPUTS Figures 24,5,6,7 display screenshots 1of the system's output screen.

EFFECTIVENESS OF 1THE SYSTEM

The suggested system uses a much more straightforward and effective method. The usage of an easy-to-use Framework makes the
system simpler. It also features significantly less complicated database setups and a more effective algorithm. Due to its platform
independence, the system is more effective.

CONCLUSION
The suggested system can be used to get people's attendance and record when they enter and leave. The system is extensively
applicable in institutions and businesses. The suggested system continuously monitors 1the entry and exit points to1 record each
student's attendances. Our pilot experiment's findings indicate that our method performs better than conventional attendance
marking methods at estimating1 attendance.

REFERENCES

1. Ashish Choudhary,Abhishek Tripathi,Abhishek Bajaj,Mudit Rathi and B.M Nandini IEEE 8/6/2012]
2. By Alan D. Moore , Python GUI Programming with Tkinter Develop Responsive and Powerful GUI Applications with
Tkinter , IEEE 4/11/2010]
3. Mrunal Aware, Prasad Labade, Manish Tambe, Aniket Jagtap, Chinmay Beldar, "Attendance Management System using
Face-Recognition", International Journal of Scientific Research in Computer Science, Engineering and Information
Technology (IJSRCSEIT), ISSN : 2456-3307, Volume 7 Issue 3, pp. 336-341, May-June 2021. Available at doi:
https://doi.org/10.32628/CSEIT217370 Journal URL: https://ijsrcseit.com/CSEIT217370
4. Thakurendra singh and V.K.Tomar in scopus journal “Performance analysis of high speed Amplifiers for low power
application with 45nm cmos technology”Harbin Gongye Daxue Xuebao/journal of Harbin Institute of
technology(2021),Page No(53) Volume No(12),197-203,ISSN:0367-6234
5. Information Science and Engineering, The National Institute of Engineering, Automatic Attendance System Using Face
Recognition. IRJET 12/05/2013]
6. Anushka Waingankar1, Akash Upadhyay, Ruchi Shah, Nevil Pooniwala, Prashant Kasambe. Face Recognition based
Attendace Mangement System using Machine Learming.
7. Python Documentation [Online].
URL:http//www.w3school.com/python/python into.asp

IJRTI2304081 International Journal for Research Trends and Innovation (www.ijrti.org) 498