2024 5th International Conference for Emerging Technology (INCET)

Karnataka, India. May 24-26, 2024

Smart Camera for Attendance Management System

Tashi Gyeltshen Janaki Kandasamy
Department of CSE-AI Department of CSE-AI
Jain Deemed-To-Be University Jain Deemed-To-Be University
Banglore 562 112, India Banglore 562 112, India
[email protected] [email protected]

Arbin Acharya Abdella Yusuf Omar

Department of CSE-AI Department of CSE-AI
Jain Deemed-To-Be University Jain Deemed-To-Be University
Banglore 562 112, India Banglore 562 112, India
[email protected] [email protected]

Abstract— The project aims to develop a smart camera-

based attendance management system using AI for automatic
In summary, the project introduces an innovative facial
attendance recording in classrooms, universities, offices, and recognition solution to enhance attendance management.
2024 5th International Conference for Emerging Technology (INCET) | 979-8-3503-6115-5/24/$31.00 ©2024 IEEE | DOI: 10.1109/INCET61516.2024.10593043

organizations. Employing computer vision algorithms for face With a focus on automation, efficiency, and adaptability, the
detection and recognition, the system captures images through system aims to provide a reliable alternative in diverse
smart cameras and updates attendance records in a database settings, contributing to the advancement of face recognition
without human intervention. The use of deep learning-based technology.
facial recognition algorithms and OpenCV for image processing
ensures accuracy. The system is scalable and customizable to
II. LITERATURE SURVEY
meet diverse institutional needs. With a user-friendly web
interface, administrators can efficiently manage attendance
Our project involves a comprehensive exploration of
reports. By eliminating manual processes, the system enhances
efficiency, reduces errors, and has the potential to revolutionize attendance management systems, each employing distinct
attendance management across educational institutions, offices, technologies to address the challenges associated with
and organizations. traditional methods. This literature survey encompasses five
notable systems, providing a holistic overview of their
Keywords— Smart cameras, Artificial Intelligence (AI), features, advantages, and potential applications.
Attendance management system, Facial recognition, Computer
vision algorithms, Deep learning, OpenCV. The NFC-based system integrates Near Field
Communication (NFC) technology with a mobile application,
I. INTRODUCTION offering an efficient attendance management solution. Each
student receives a unique NFC tag during enrolment,
The project tackles attendance management challenges in simplifying the attendance process through touch or
educational institutions, proposing an automatic student movement on the lecturer's mobile phone. The embedded
attendance system using facial recognition technology. This camera captures students' faces, enhancing validation.
solution eliminates the need for student input and addresses Despite its simplicity, concerns arise regarding potential
issues associated with traditional biometric methods[8]. fraud and inconvenience for lecturers, especially when
Utilizing the Voila and Jones algorithm for efficient face relying on personal smartphones[1].
detection, the system aims to streamline administrative
processes, saving time and reducing workload[7]. Utilizing face recognition technology, this system
prioritizes accuracy and fraud prevention. Attendance is
Motivated by the pivotal role of attendance management marked on a secure server, ensuring data confidentiality. The
in schools, the project emphasizes the efficiency of an face detection algorithm is enhanced through a skin
automated system that offers real-time tracking, analysis of classification technique. However, the system's non-
student absences, and eliminates redundant data entry. The portability limits its application to staff attendance, and
main objective is to develop a facial recognition attendance potential performance issues under suboptimal conditions
system applicable in various environments, including need consideration [2].
schools, workplaces, factories, and homes[9].
This system leverages fingerprint recognition and
Despite being in a developing stage, the technology holds microcontrollers for efficiency and reliability. The
promise for workforce monitoring without human integration with LabVIEW facilitates data visualization.
intervention. Environments identified for face recognition While offering accuracy and fraud prevention, its reliance on
use include schools for class attendance monitoring, a PC makes it non-portable. Limited accessibility to
workplaces, factories, and homes for diverse applications attendance information for parents is a notable drawback, and
such as monitoring children and pets[10]. potential issues may arise in detecting fingerprints under
certain conditions[3].

979-8-3503-6115-5/24/$31.00 ©2024 IEEE 1

Authorized licensed use limited to: Anglia Ruskin University. Downloaded on August 12,2025 at 14:52:54 UTC from IEEE Xplore. Restrictions apply.
 The RFID-based system employs tags and a reader for This continuous loop ensures real-time, accurate attendance
attendance tracking, providing information accessibility management, making the system efficient and reliable.
through a web portal. However, non-portability due to the
need for a connected PC and concerns about accuracy in
uniquely identifying students using RFID tags are notable
limitations[4].

Incorporating Convolutional Neural Network (CNN)

architecture, this system focuses on improved accuracy and
automation. With a web application for online attendance
posting, it stands out for its potential advancements.
However, complexity in development, cost considerations,
scalability issues, and privacy concerns must be carefully
addressed[5].

This extensive literature survey provides a nuanced

understanding of various attendance management systems,
shedding light on their distinctive technological attributes, Fig. 1. Flowchart of Face Detection Algorithm
advantages, and limitations. As our project aims to contribute
The flowchart represents a face recognition attendance
to the advancement of attendance management, these insights system. It captures frames from a webcam, detects and
serve as valuable reference points for informed decision- recognizes faces, and updates attendance data in real-time.
making and system design. The program integrates with Firebase for data storage and
retrieval. The system operates in a continuous loop,
The explored NFC-based system showcases simplicity processing frames, comparing face encodings with known
and efficiency but calls attention to potential vulnerabilities encodings, retrieving student information from the database,
and inconveniences associated with personal smartphone and displaying attendance details.
usage for attendance tracking. On the other hand, the face
recognition system emphasizes accuracy and security, with B. Alogrithm 2: Latecomer Tracking
server-based marking and enhanced detection algorithms, but
its limitations in portability may restrict its applicability to The attendance system incorporates an advanced
staff attendance[2,4]. algorithm to track and manage late arrivals during events.
Upon detecting a person's face, the system captures a
III. PROPOSED METHODOLOGY timestamp. It then determines the scheduled start time for the
event, calculating the time difference between the recorded
In our collective approach, we systematically integrate timestamp and the scheduled commencement. To establish
cutting-edge technologies to actualize the objectives tardiness, a configurable latecomer threshold is defined, often
articulated in the project's introduction. Employing deep set at, for instance, 5 minutes. Comparing the calculated time
learning algorithms for facial recognition, we harness the difference with the established threshold, the system flags
capabilities of the OpenCV library to handle image individuals surpassing this limit as latecomers. Subsequently,
processing intricacies[7]. The core of our methodology lies in the system records pertinent details, such as the latecomer's
the deployment of smart cameras strategically positioned name or identification. To enhance administrative oversight,
within the premises, capturing individuals upon entry. Our the system generates comprehensive reports or notifications
computer vision algorithm then meticulously analyses these highlighting the late arrival status of marked individuals. This
images, facilitating robust face detection and recognition. meticulous process ensures that late arrivals are accurately
The identified faces undergo cross-referencing with a identified and recorded, contributing to effective attendance
comprehensive database, ensuring automated and accurate management. The incorporation of configurable parameters,
attendance marking. The user-centric design extends to a such as the latecomer threshold, provides adaptability to
web-based interface, enhancing accessibility and efficiency diverse event scenarios. Overall, the system's automation
for administrators managing attendance records. This significantly streamlines the identification and reporting of
collaborative methodology ensures the seamless integration latecomers, facilitating efficient event management.
of AI- driven smart cameras for attendance automation,
The flowchart depicts the latecomer management process
showcasing its practicality and effectiveness across various in an attendance system. It initiates with timestamp capture
settings[9]. upon detecting a face, determining the scheduled start time,
and calculating the time difference. A predefined latecomer
A. Alogrithm 1: Face Recognition threshold identifies late arrivals. Details are recorded, and
reports/notifications signify late arrival status, ensuring
The algorithm seamlessly integrates Firebase and
efficient latecomer tracking.
webcam functions, capturing frames, resizing, and converting
them for face detection. Utilizing face_recognition, it
compares encodings, identifying students and retrieving their
information from Firebase. Attendance updates, based on
elapsed time, dynamically display on the background image.

2
Authorized licensed use limited to: Anglia Ruskin University. Downloaded on August 12,2025 at 14:52:54 UTC from IEEE Xplore. Restrictions apply.
 Overall Flowchart

Fig. 2. Flowchart for Latecomer Tracking Algorithm

C. Alogrithm 3: Face Alignment

The face alignment process commences with an input face
image, utilizing a pre-trained model to detect facial
landmarks and pinpoint essential features like eyes, nose, and
mouth. Guided by specific criteria or reference points, the
algorithm determines the desired alignment for the face.
Calculating transformation parameters, including rotation
and translation, facilitates the realization of the defined
alignment. The algorithm applies these parameters,
transforming the image and presenting the aligned face for
further processing or analysis. Significantly, this face
alignment technique ensures consistent and accurate face Fig. 4. Overall Flowchart
orientation and positioning in images, optimizing tasks such
as facial recognition. Emphasizing its versatility, face The flowchart illustrates an algorithm for facial landmark
alignment is instrumental for future purposes, laying the detection and visualization. It begins with an input image of a
groundwork for ongoing advancements in facial analysis, face, which is then converted to a simplified black-and- white
recognition, and diverse image processing applications. version. The algorithm proceeds to detect the presence of a
face in the image. If a face is found, the algorithm utilizes a
pre-trained model to locate and identify the positions of key
facial landmarks such as the eyes, nose, and mouth. The
coordinates of these landmarks are extracted and marked on
the image, which is then displayed with the visual
representation of the facial landmarks.

IV. RESULT ANALYSIS

Fig. 5. Output of source code

This figure shows the items that are displayed when the
code is being executed and a face has been recognized and
matched with the face from the database. What is happening
here is that the time of the recognition (attendance time), the
student’s major, name, starting year, total updated attendance
and year of study are first being outputted in the console.
Fig. 3. Flowchart of Face Alignment Algorithm

3
Authorized licensed use limited to: Anglia Ruskin University. Downloaded on August 12,2025 at 14:52:54 UTC from IEEE Xplore. Restrictions apply.
 Figure 7 shows the output from the third algorithm that is
the Face Alignment Algorithm The main functionality of this
algorithm is to capture face landmarks (especially eyes) by
referencing the pretrained model and store it for future use
cases.
In the figure 8, we see the real time database before and
after a face has been detected and a match has been found.
Before the match is found, the attendance of the student with
USN 20btrca000 remain 9. After the face has been detected
and the match with that particular student had been found, the
attendance has been updated to 10 from previously being 9.

V. CONCLUSION

Our collaborative endeavour in crafting an advanced

attendance management system showcases the
transformative power of artificial intelligence (AI) and smart
camera technologies. The incorporation of deep learning
algorithms, particularly leveraging OpenCV for face
detection and Convolutional Neural Network (CNN)
Fig. 6. Screenshot of the face detection and attendance marking process architecture for precise face recognition, has propelled the
when the code is executed. system to new heights of accuracy and efficiency across
diverse environments such as classrooms, offices, and
This figure shows how the display of the attendance looks
universities[5,7].
when the code is being executed. In the above picture, we see
that the face has been detected and it has found a match from
Our face alignment algorithm, rooted in detecting facial
the database. In the second picture, we see that because the
landmarks, ensures a consistent and accurate orientation of
match has been found, the attendance has been marked.
faces, elevating the reliability of facial analysis and
recognition. This strategic amalgamation of technologies not
only eradicates the need for manual attendance marking but
also optimizes administrative processes by saving time and
minimizing errors[9].

Beyond its immediate application, our system sets the

stage for future developments. The integration of Firebase for
data management, latecomer tracking, and dynamic reporting
underscores the adaptability and scalability of our solution.
Emphasizing face alignment reflects our commitment to
Fig. 7. Output of Face Alignment future-proofing the system, ensuring its relevance and
effectiveness amidst evolving technological landscapes[7].

In essence, this project can not only showcase the current

patterns of the AI-based attendance management systems, but
also can pioneer the prospective trends closely related to the
development, and necessary to the future of time-keeping
systems. Using the most advanced technologies, we have
created an offering that is a good fit for the immediate needs
of many institutions and is adaptable and able to embrace the
probable future trends in the attendance tracking and facial
recognition technology sector[7,9].

REFERENCES
[1] Bhise, A., Khichi, A., Korde, N., & Lokare, A. (2015). Attendance
system using NFC technology with embedded camera on mobile
device. International Journal of Advanced Research in Computer
Engineering & Technology, 4(11), 4127-4131.
[2] SenthamilSelvi, N., Chitrakala, S., & Antony Jenitha, J. (2014). Face
recognition based attendance marking system. International Journal of
Advanced Research in Computer Science and Software Engineering,
4(3), 168-172.
[3] Kumar Yadav, A., Singh, V., Pujari, J., & Mishra, S. (2015). Fingerprint
based attendance system using microcontroller and LabView.
International Journal of Advanced Research in Computer Science and
Fig. 8. Screenshot of Realtime Database Software Engineering, 5(3), 1225-1229.

4
Authorized licensed use limited to: Anglia Ruskin University. Downloaded on August 12,2025 at 14:52:54 UTC from IEEE Xplore. Restrictions apply.
[4] Hussain, A., Dugar, D., Deka, J., & Hannan, M. A. (2014). RFID based [7] Li, S. Z., & Jain, A. K. (Eds.). (2011). Handbook of face recognition
student attendance system. International Journal of Advanced Research (2nd ed.). Springer.
in Computer Engineering & Technology, 3(10), 3130-3135. [8] Zhang, D., & Jain, A. K. (2005). Biometrics: Personal identification in
[5] Syam, V., Gangula, R., Rahul, K., & Singh, A. (2018). Smart networked society. Springer.
attendance management system based on face recognition using CNN. [9] Dey, N., Ashour, A. S., & Shi, F. (2018). Smart environments:
International Journal of Engineering Research & Technology, 7(6), Technology, protocols, and applications. CRC Press.
269-275.
[10] Jain, A. K., Ross, A., & Pankanti, S. (2006). Biometrics: A tool for
[6] Sharma, P., Kumar, A., Gupta, R., & Verma, D. (2016). Automated information security. IEEE Transactions on Information Forensics and
attendance system using facial recognition. International Journal of Security, 1(2), 125-143.
Computer Applications, 147(10), 31-34.

5
Authorized licensed use limited to: Anglia Ruskin University. Downloaded on August 12,2025 at 14:52:54 UTC from IEEE Xplore. Restrictions apply.