2023 7th International Conference On Computing, Communication, Control And Automation (ICCUBEA)

Pimpri Chinchwad College of Engineering (PCCOE), Pune, India. Aug 18-19, 2023

Multi-Facial Automated Attendance System using

Haar Cascade, LBPH, and OpenCV-Based Face
Detection and Recognition
Snehal Rathi Snehal Kapadnis Gayatri Sonar
2023 7th International Conference On Computing, Communication, Control And Automation (ICCUBEA) | 979-8-3503-0426-8/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICCUBEA58933.2023.10392248

Vishwakarma Institute of Information Vishwakarma Institute of Technology Software Engineer, Aress Software,
Technology Pune, India Nashik, India
Pune, India [email protected] [email protected]
[email protected]

Gaurav Sonar Sushma R. Vispute

Vishwakarma Institute of Information Pimpri Chinchwad college of Engineering
Technology Pune, India
Pune, India [email protected]
[email protected]

Abstract- In the 21st century, everything around us is II. PROBLEM STATEMENT

dependent upon technology to make our lives much easier.
Daily tasks are increasingly being computerized. Nowadays Lecturers that use the current attendance system find
more people prefer to do their work electronically. To the issues such as manually recording student attendance,
extent of our knowledge, students' attendance at some messing up attendance papers, dishonesty, lost time, and high
universities is manual. The attendance management system loss rates. It is a difficult process that takes time and involves
computerizes the traditional attendance process. Utilizing a lot of paperwork. So, to address these issues and prevent
OpenCV (Open-source Computer Vision) and some face mistakes, we recommend computerized procedure by
detection and recognition algorithms, we propose a system to offering a software program that automatically tracks and
measure student attendance in classes. By using face detection monitors student attendance without the involvement of
algorithms and face recognition algorithms, we will keep track teachers.
of how many people are present in the class and attendance of
each one. This project intends to provide a system that III. LITERATURE REVIEW
automates and streamlines the process of collecting and
The proposed system addresses the issue of proxy
monitoring student attendance using facial recognition
technology. To find, identify, and validate collected faces, pre-
attendance and students being marked as present even when
processing methods like the Haar Cascade approach will be they are not present physically. The system makes use of live
used. To make the attendance process quicker and more video stream and from the video frames are extracted using
accurate, we aim to offer a method. OpenCV library. It uses dlib library to detect the faces and
recognize them. After these, by comparing the actual faces
Keyword: attendance management system, computer vision, with the database which contains students' faces, the
face detection, face recognition, Haar Cascade approach, attendance of respective student is marked successfully [1].
I. INTRODUCTION The Proposed system is implemented in four phases:
Image Capturing, Image Segmentation from group image
The procedure for recording student attendance has and based on that image Face Detection will be done [8],
evolved and changed during the last ten years. The need to Face comparison will be done against database, Lastly
automate, simplify, accelerates, and save time and effort is Attendance will get update accordingly. The objectives are to
what propels this progress. The teacher calls out for each save time and make system simple and easy to use [2].
student and individual to mark their attendance while using
the manual attendance method, even though attendance The Attendance Management Using Facial Recognition
systems are presents everywhere. Even if it is a time- enrolls the faces of student subject wise into the database
consuming process, there are also chances of having manual against the subject ID and Name[12]. The system then allots
errors. Also, one of the main problems teachers could have in attendance to the recognized faces in the database [3].
class is low attendance. In this project, we try to cut down on Several techniques are discussed for detecting the face,
idling time, get rid of mistakes, and automate the procedure. such as the Ada-Boost algorithm for detection of the faces;
Our method uses facial recognition technology to track the Viola Jones algorithm for face detection, SNOW
attendance using a high-resolution camera that can detect and Classifier method and the SMQT features, and Local Binary
identify the faces. It then compares the identified faces with Pattern (LBP) are used [4].
students face images recorded in a database. An individual's
attendance is noted in the attendance database once their face Students or staff must wait for the completion of the
matches a saved photograph. Consequently, the focus of our previous members to prevent proxy issues and maintain
project will be on online student attendance. In other words, attendance effectively [9]. The proposed system detects and
the objective of this paper is to create a system which enables recognizes iris patterns by having the student stand in front
professors to take student attendance in a accurate automated of a camera. In order to detect the iris, some algorithms are
manner. used, such as Gray Scale Conversion, Rectangular Filter, and
Skin Pixel Detection [5].

979-8-3503-0426-8/23/$31.00 ©2023 IEEE 1

Authorized licensed use limited to: Modern Education Society's Wadia College of Engineering. Downloaded on July 22,2024 at 10:20:50 UTC from IEEE Xplore. Restrictions apply.
 Face recognition has been improved through numerous V. METHODOLOGY
algorithms and techniques, but Deep Learning is being
applied here [11][13]. This tool converts the frames of the A. Face detection
video into images in order to identify the face of the student In many applications, face detection is a technique for
so that their attendance can be automatically reflected in the recognizing human faces in digital images. At the process of
attendance database [6]. face detection, faces will be found in captured photographs
so that they can be compared. The Viola-Jones method is
Face detection rate is not high enough to allow us to used to detect faces.
estimate attendance precisely based on individual face
recognition results[10]. In this paper author propose the Haar Cascade Algorithm: This is commonly used
method to estimate attendance precisely using continuous algorithm for object detection. This algorithm's key
observation which improves the accuracy of attendance[14]. characteristic is that detection happens quickly while training
The system uses raspberry pi with OpenCV library. Thus, the takes a while. Generating the integral image first can greatly
attendance approximation performance will be improved due improve the performance of this technique.
to uninterrupted observation [7].
x Haar features.
A. Existing
stiing System:
x Integral Image
Lecturers still utilize the manual attendance system,
which involves calling out students' names or circulating an x Cascading
attendance sheet among them so that each student can sign it
next to their name. Both methods take a lot of time and have
large error margins. There are many issues with collecting
attendance in this manner traditionally, as students
occasionally fill in for friends or teachers who are absent.
B. Proposed System:
By taking a picture of the entire class and uploading it to
a system that detects and recognizes the students' faces,
facial recognition technology is used to track attendance. The
system then matches the identified face with student face
Fig. 1. Haar Feature on Training Image
photographs that are kept in the database. Based on the result
of comparison, attendance of that student for respective
subject is marked. 1) Face Recognition
The face recognition algorithm oversees determining the
IV. OBJECTIVE features that best define the image once the facial images
have been retrieved, cropped, scaled, and typically converted
Our main objective is to support the teachers in their
to grayscale. Face recognition algorithms come in a variety
efforts to track and control student attendance and
of forms, such as Eigen Faces, LBPH (Local Binary Patterns
absenteeism by streamlining and organizing the process. In
Histograms), and Fisher Faces. For the suggested system,
addition, using this system,
LBPH is the most appropriate of these algorithms.
x Teachers can keep track of defaulters and take B. LBPH Algorithm:
suitable action at the correct time.
A straightforward yet very effective texture operator, the
x Students can keep track of their attendance. Local Binary Pattern (LBP), labels pixels a picture by
thresholding the area around each pixel. Steps involved in
x By offering an automatic and trustworthy attendance
system which makes use of facial recognition x Radius, Neighbors, Grid Parameter Xand Y.
technology, manual process mistakes can be
reduced. x Practicing the LBP operation.

x It boosts privacy and security so that students can't x Extracting the histograms.
introduce themselves or their friends when they x Training the algorithm
aren't there.
x Using face recognition
x Being adaptable and having the flexibility to change
attendance recordsfor lectures. C. OpenCV
Features One of the objectives of OpenCV is to provide a simple-
to-use computer vision infrastructure that allows the
x The device automatically records attendance and development of sophisticated vision applications. One of
retains the faces that are detected. OpenCV's goals is to provide a computer vision
infrastructure that is easy to use so that people can build
x Allow authorized access. sophisticated vision applications quickly. Face detection,
x Ease of use. feature extraction, and face recognition are the three stages of
a facial recognition system, as shown in Figure.
x Several faces detected at a time.

2
Authorized licensed use limited to: Modern Education Society's Wadia College of Engineering. Downloaded on July 22,2024 at 10:20:50 UTC from IEEE Xplore. Restrictions apply.
 attendance for recognized faces. Also, users can view the
attendance of present students from View Attendance tab.
So, like this, it will work very efficiently and in user friendly
manner.
VII. ARCHITECTURE DIAGRAM

Fig. 2. Structure of face recognition system

Fig. 3. Architecture Diagram
VI. IMPLEMENTED SYSTEM
We propose a system that finds and recognizes faces VIII. RESULTS AND DISCUSSION
using the Haar Cascade object detector method. This may be A. Image Upload:
accomplished using Open CV's in-built Haar cascades.
Trained faces are first entered into a database, and the trained Firstly, we need to register students into the database and
photos are then compared (trained images are initially upload image while registering student. The image will store
available in the database indicates the linked folks to the in folder with name of student ID.
specific college or organization). That will result in
attendance to the people who have been identified after
comparison.
A. Faculty Module:
In this section, there is functionality like add faculty, in
which we can register new faculty members in our domain
by taking some required details and store these details in
our database. There will be options also like department,
class, or division wise we can show all the faculty members.
B. Student Module:
The student module works like the faculty module. It will
also have functionality of register student, modify details of
student, displaying all the students present in domain. But Fig. 4. Primary Faces Registration
while registering students we are taking input images of
students which will be work for our attendance module. B. Face Detection:
C. Subject Module: The image is used as an input for face detection, which
looks for any faces. Once the face is located, the image is
In this Subject tab, there will be subject registration and cleaned up for easier recognition. Facial recognition can be
subject allocation. Here we can add subject related details achieved by implementing the Haar Cascade algorithm.
and then allocate subjects to the faculty member. From this Datasets are trained using Haar Cascade to obtain unique
faculty will understand the subject or class for which they features of their face (eyes, noses, ears, lips). It works by
need to mark the attendance. According to that faculty will giving us most important parts of the face for detection
mark the attendance of different subject lectures. through Haar cascade algorithms.
D. Attendance Module: C. Face Matching
The attendance module is the most important part or The LBPH algorithm (Local Binary Pattern Histogram) is
main objective of this project. So, in this, we are marking the used to compare the captured images with the stored images
attendance of students which are present at the time of in the database to recognize the student. Multiple student
lecture or any class. Here we must upload the images of faces are detected by the system and according to face
students. Once the model is trained using algorithms system recognition attendance will be marked and stored into
detects and recognizes the face. After comparing the image database. A single instance will be able to handle at least 15
with the database's trained images, the algorithm records faces.

3
Authorized licensed use limited to: Modern Education Society's Wadia College of Engineering. Downloaded on July 22,2024 at 10:20:50 UTC from IEEE Xplore. Restrictions apply.
 D. Fetch Attendance REFERENCES
Students or Faculty can fetch attendance by proving [1] Hl, D.G., Vishal, K., Dubey, N.K. and Pooja, M.R., 2020. Face
appropriate data like Department, Date and for which subject recognition-based attendance system. International Journal of
Engineering Research & Technology (IJERT), 9(6).
they want to fetch attendance of present students.
[2] Gomes, C., Chanchal, S., Desai, T. and Jadhav, D., 2020. Class
attendance management system using facial recognition. In ITM Web
of Conferences (Vol. 32, p. 02001). EDP Sciences.
[3] Bharadwaj R., Rao T., R. 2020. Attendance Management Using
Facial Recognition. International Journal of Innovative Technology
and Exploring Engineering (Vol-8 Issue-6,).
[4] Gupta, V. and Sharma, D., 2014. A study of various face detection
methods. International Journal of Advanced Research in Computer
and Communication Engineering, 3(5), pp.6694-6697.
[5] Singh D., Kusuma M., Mounika K. 2019. IOT Based Automated
Attendance with Face Recognition System. International Journal of
Innovative Technology and Exploring Engineering (Vol-8, Issue-
6S4).
[6] Nandhini, R., Duraimurugan, N. and Chokkalingam, S.P., 2019. Face
recognition-based attendance system. International Journal of
Fig. 5. Fetching the attendance fromdatabase. Engineering and Advanced Technology (IJEAT), 8(3S)
[7] Katara, A., Kolhe, M.S.V., Zilpe, M.A.P., Bhele, M.N.D. and Bele,
Advantages M.C.J., 2017. Attendance system using face recognition and class
monitoring system. International Journal on Recent and Innovation
x This system can be applied to secured areas and Trends in Computing and
organizations for security purposes. [8] Communication, 5(2), pp.273-276
x The software automatically records attendance while [9] Kalyankar, P.A., Mulani, A.O., Thigale, S.P., Chavhan, P.G. and
Jadhav, M.M., 2022. Scalable face image retrieval using AESC
storing the observed faces. technique. Journal Of Algebraic Statistics, 13(3), pp.173-176
x Technology offers users convenience and security. [10] Kaur, H., & Singh, A. (2020). An intelligent attendance management
system using face recognition. In 2020 7th International Conference
x That helps them save time and effort. on Signal Processing and Integrated Networks (SPIN) (pp. 156-161).
IEEE.
Disadvantages [11] Harshad Patil, Heena Khan, Saquib Ansari, Saimohan Sahu., 2021.
Automatic endance System Using Face Recognition. In International
x The system doesn't recognize faces well in low light, Journal of Advanced Research in Science, Communication and
which could lead to erroneous results. Technology (IJARSCT) (Volume 4, Issue 1) SSN (Online) 2581-
9429.
x It is only capable of detecting close up faces. [12] S. R. Rathi and Y. D. Deshpande, "Embedding Affect Awareness into
Online Learning Environment using Deep Neural Network," 2019 5th
IX. CONCLUSION International Conference On Computing, Communication, Control
And Automation (ICCUBEA), Pune, India, 2019, pp. 1-6, doi:
Hence, a system that is trustworthy, secure, quick, and 10.1109/ICCUBEA47591.2019.9128811.
efficient has been devised to replace a manual and unreliable
[13] Ms. Dyanila Neeta Ferrao, Ms. Kusuma D R, Mr. Manoj Kammar,
attendance system. Face detection using Haar-Cascades Mr. Pradeep N M, Mr. Vishal Hegde.,2022. Student Attendance
performed incredibly well. A cost-effective face recognition System Using Face Recognition. IJRTI. (Volume 7, Issue 6), ISSN:
platform can be created by combining LBPH and Haar- 2456-3315
Cascades. For better outcomes in the management of [14] S. R. Rathi and Y. D. Deshpande, "Embedding Affect Awareness
attendance, this approach can be put into place. The into Online Learning Environment using Deep Neural
technology will save time, lessen the amount of work that Network," 2019 5th International Conference On Computing,
Communication, Control And Automation (ICCUBEA), Pune, India,
needs to be done by the administration, replace paper 2019, pp.1-6,doi: 10.1109/ICCUBEA47591.2019.9128811.
documents with digital ones, and cut down on the number of [15] S. Rathi, Y. Deshpande, S. Nagaral, A. Narkhede, R. Sajwani and V.
people needed for the job. Takalikar, "Analysis of User’s Learning Styles and Academic
Emotions through Web Usage Mining," 2021 International
Conference on Emerging Smart Computing and Informatics (ESCI),
Pune, India, 2021, pp. 159-164, doi:
10.1109/ESCI50559.2021.9397037.

4
Authorized licensed use limited to: Modern Education Society's Wadia College of Engineering. Downloaded on July 22,2024 at 10:20:50 UTC from IEEE Xplore. Restrictions apply.