Attendance Management System using Face Recognition

A Project Report

submitted in partial fulfillment of the requirements

of

AICTE Internship on AI: Transformative Learning

with
TechSaksham – A joint CSR initiative of Microsoft & SAP

by

Nandhidakavi A L, [email protected]

Under the Guidance of

Aditya Prashant Ardak

 ACKNOWLEDGEMENT

I am immensely grateful for the guidance, encouragement, and unwavering support

that I have received throughout the course of this project. This accomplishment would not
have been possible without the contributions of several individuals and organizations, to
whom I owe my deepest gratitude.

First and foremost, I wish to express my heartfelt thanks to my esteemed guide, Aditya
Prashant Ardak for their exceptional mentorship and dedicated guidance. Their insightful
advice and constructive feedback have been invaluable throughout every stage of this project.
The patience and expertise demonstrated by my guide in addressing challenges, providing
innovative solutions, and steering me in the right direction have been the cornerstone of this
successful endeavor. Their belief in my abilities and continuous motivation inspired me to
push the boundaries of my potential and achieve excellence in this work.

I am also profoundly grateful to TechSaksham for providing such an enriching platform to

explore and implement innovative ideas in the field of artificial intelligence. The
transformative learning experience and access to invaluable resources provided through this
initiative have significantly enhanced my technical knowledge and professional
development. The internship offered me a unique opportunity to translate theoretical
concepts into practical applications, and I deeply appreciate the vision of TechSaksham in
empowering young minds like me.
 ABSTRACT

This report presents the development of an AI-driven Face Recognition-Based Attendance

Management System designed to automate attendance tracking. Traditional attendance
methods, such as manual roll calls and sign-ins, are inefficient and prone to errors, especially
in scenarios involving large-scale participation. To address these challenges, the proposed
system employs real-time face detection and recognition using the face recognition library
and provides an intuitive, user-friendly interface built with Tkinter.

The system offers key functionalities, including real-time attendance marking with color-
coded feedback, student registration, and attendance logs with timestamps for auditing. The
solution emphasizes accuracy and efficiency by leveraging advanced facial encoding
techniques, ensuring robust performance under moderate lighting and angle variations.

The project also addresses key limitations of existing systems, such as scalability issues,
environmental sensitivity, and privacy concerns. Experimental evaluations demonstrate the
system's reliability and suitability for small- to medium-scale setups, such as classrooms and
small offices. Future enhancements will focus on integrating cloud-based storage, improving
recognition in challenging conditions, and ensuring compliance with privacy standards.

This report outlines the methodology, implementation, and potential impact of the system,
showcasing its ability to streamline the attendance process and reduce administrative
burdens while enhancing accuracy and transparency.

.
 TABLE OF CONTENT

Abstract ........................................................................................................................ I

Chapter 1. Introduction ......................................................................................... 1

1.1 Problem Statement................................................................................. 1
1.2 Motivation ............................................................................................. 1
1.3 Objectives .............................................................................................. 2
1.4. Scope of the Project .............................................................................. 2
Chapter 2. Literature Survey ................................................................................ 3
2.1 Review relevant literature ...................................................................... 3
2.2 Existing Models, Techniques, and Methodologies ................................ 4
2.3 Limitations in Existing Systems ............................................................ 4
Chapter 3. Proposed Methodology ........................................................................ 6
3.1 System Design ....................................................................................... 6
3.2 Requirement Specification..................................................................... 8
Chapter 4. Implementation and Results ............................................................... 9
4.1 Snap Shots of Result.............................................................................. 9
4.2 GitHub Link for Code ......................................................................... 12
Chapter 5. Discussion and Conclusion................................................................ 13
5.1 Future Work ......................................................................................... 13
5.2 Conclusion ........................................................................................... 15
References .................................................................................................................. 16
 LIST OF FIGURES

Page
Figure No. Figure Caption
No.
Interface for the Face Recognition-Based Attendance System
Figure 1 9
Add the student details after selecting student details from the main
Figure 2 10
menu, and click save a prompt window will show up if successful.
Success! Data is saved in the local database and can be visualized in the
Figure 3 10
table at right.
Figure 4 Attendance Taken 11
Figure 5 See the saved data by importing the attendance csv 12
 CHAPTER 1
Introduction

1.1 Problem Statement:

Traditional attendance recording methods, such as manual roll calls or signing attendance sheets,
are time-consuming and error-prone. Inaccuracies arise due to human errors, proxy attendance, or
lost records, leading to inefficiencies in tracking and managing attendance. These challenges
become more significant in large-scale setups like educational institutions, corporate offices, and
events where monitoring hundreds or thousands of attendees can be overwhelming[1].

An automated attendance system based on face recognition offers a seamless, accurate, and
tamper-proof solution. By leveraging the unique biometric features of individuals, this system
ensures attendance is recorded efficiently, eliminating issues [2]

Why is this significant?

Efficiency: Saves time compared to manual methods [3]
Accuracy: Reduces human error in attendance marking.
Security: Prevents proxy attendance.
Scalability: Suitable for large-scale implementations.

1.2 Motivation:

This project was chosen due to the increasing need for automated solutions in attendance
management across various domains. As artificial intelligence (AI) and biometrics use grows,
leveraging these technologies for attendance systems can provide a reliable and innovative
solution to a common yet critical problem.[4]

Potential Applications and Impact:

1. Educational Institutions: Automates attendance in classrooms or exams, ensuring fairness

and reducing workload on staff. [1]
2. Corporate Offices: Monitors employee attendance accurately without requiring
additional hardware like fingerprint scanners.[4]
3. Events and Conferences: provide efficient attendee verification and reduce
registration delays.[5]
4. Healthcare and Public Sector: Tracks attendance in critical fields requiring precise
monitoring for compliance and accountability.[5]

1
1.3 Objective:

The goal of this project is to design and implement a Face Detection Attendance System to
automate attendance processes..[6]

Specific Objectives:[5]
1. Utilize facial recognition technology for real-time attendance marking.
2. Develop a user-friendly interface for interaction and management of attendance
records.[7]
3. Provide functionalities such as logging attendance, displaying registered faces, and
maintaining timestamped logs for auditing.[3]
4. Ensure the system is scalable, secure, and reliable[6]
5. Minimize errors and eliminate the possibility of proxy attendance.

1.4 Scope of the Project:

Scope:
The system targets small- to medium-scale setups such as classrooms, small offices, or events. It
provides real-time attendance tracking, face registration, and log management.

1. Users can take attendance, view logs, and add new students using a simple interface.[7]
2. Logs include timestamps for auditing purposes[6]
3. A webcam or external camera serves as the only hardware requirement, making it cost-
effective.[3]

Limitations:

1. Environment Dependency: Requires well-lit conditions for accurate face

recognition.[6]
2. Database Scalability: Limited by local storage; cloud integration may be needed for larger
setups. [8]
3. Privacy Concerns: Proper security measures are necessary to prevent misuse of biometric
data. [4]
4. Hardware Dependence: Requires a functioning camera and a compatible system for
smooth operation. [5]

2
 CHAPTER 2
Literature Survey

2.1 Review relevant literature

The integration of face recognition technology in attendance management systems has been
an area of active research, with several studies investigating the application of machine
learning and computer vision techniques for automating and enhancing the attendance process.

1. Real-Time Face Recognition for Attendance Tracking:

One of the most influential works in this domain is the study by Gupta et al., which employs
deep learning techniques such as Convolutional Neural Networks (CNNs) for real-time face
recognition and attendance management. Their system, developed for educational institutions,
showed improvements in both efficiency and accuracy compared to traditional attendance
methods [1]. The study also highlighted scalability concerns, especially when handling large
datasets, which can affect the recognition speed and accuracy under varying conditions.

2. FaceNet for Attendance:

The work by Singh et al. discusses the application of FaceNet for embedding face vectors and
using them to match faces for attendance. The study emphasizes the advantages of using a
deep learning approach, achieving high accuracy in face recognition, but also discusses the
computational overhead involved when implementing the system in real-time scenarios,
particularly on resource-constrained devices [3].

3. Combination of Face Recognition and IoT:

Some studies combine face recognition with Internet of Things (IoT) devices for smarter
attendance systems. An example is the research by Sharma et al., where facial recognition is
integrated with smart systems to not only track attendance but also enable seamless data
storage in cloud databases, making the system more scalable and easier to access remotely [3].
However, their approach faces challenges with data security and privacy concerns, which are
common in biometric systems.

3
 4. Haar Cascade and Local Binary Patterns (LBP):
Several systems, including the one by Kumar et al., use traditional methods like Haar Cascades
for face detection and Local Binary Patterns (LBP) for feature extraction. While these
models are lightweight and faster for real-time applications, they are less robust under poor
lighting or when faces are partially occluded, as noted by the authors. Despite this, they offer a
solution for small-scale systems requiring low computational power [4] .

5. Challenges with Traditional Models:

A major limitation noted in the literature is the challenge of handling complex backgrounds
and occlusions in real-world scenarios. For example, models based on Support Vector
Machines (SVM), as discussed in research by Hati et al., can produce false positives when
deployed in fixed or controlled environments, highlighting the need for more adaptive, neural
network-based models for robust performance [5].

2.2 Existing Models, Techniques, and Methodologies

1. Haar Cascade Classifiers: Traditional face detection method offering lightweight operation
but limited robustness under varying lighting and angles[1].
2. Deep Learning Models (e.g., FaceNet, OpenFace): Provide high accuracy through deep
embeddings but are resource-intensive and require significant computational power for real-
time performance [3].
3. YOLO-Based Detection: Faster models like YOLO have been used for real-time applications
but need optimized preprocessing pipelines for small-scale attendance systems.[1]

2.3 Limitations in Existing Systems

1. Scalability Issues: Current systems face performance degradation as the database grows,
impacting recognition speed.

4
2. Environmental Sensitivity: Accuracy reduces significantly under poor lighting, extreme
facial angles, or partial occlusions.
3. Privacy Concerns: Some implementations inadequately address data security and consent,
leading to ethical challenges.

How This Project Addresses the Gaps

1. Real-Time Performance: Utilizes efficient libraries like face_recognition and
lightweight UI design through Tkinter for optimized performance on standard hardware.
2. Enhanced Accuracy: Employs post-detection refinements, such as color-coded status
for user-friendly operation.
3. Privacy-First Design: Data is securely stored in Firebase with minimal personal data
collection.
4. User Experience: Simplifies operation with intuitive interface options like
attendance, logs, and registration.

5
 CHAPTER 3
Proposed Methodology

The proposed methodology outlines the system design and implementation strategy for the face-
recognition-based attendance management system. It ensures real-time operation, user-friendly
interaction, and secure data handling.

3.1 System Design

The system design integrates several interconnected modules to ensure smooth functionality:

1. Face Detection Module:

o Uses the face_recognition library powered by dlib to detect human faces in real-time.
o Ensures high accuracy in identifying faces under standard lighting and moderate
angle variations [1], [5]
2. Recognition and Verification Module:
o Matches detected faces with a pre-registered database of face encodings.
o Displays the recognition status in a color-coded format:
• Red: Face detected but not recognized.
• Green: Face successfully recognized [5]
3. Attendance Marking Module:
o Updates the attendance database upon successful recognition.
o Associates the recognized name with a timestamp for logging purposes.
4. Log Management Module:
o Maintains a history of recognized faces with timestamps for record-keeping and
validation [9].
5. User Interface (UI) Module:
o Built using Tkinter for simplicity and accessibility.
o Features intuitive options:
• Taking Attendance
• Viewing Registered Face
• Checking Logs
• Registering New Students [4

6
Figure 1: System Workflow for Face Recognition-Based Attendance
Management System

7
3.2 Requirement Specification

3.2.1 Hardware Requirements:

o Camera/Webcam: For capturing real-time images.[2]
o Processing Unit: Dual-core processor or higher for smooth computation.
o RAM: Minimum 4GB to handle real-time processing efficiently.
o Storage: Adequate space for storing logs and face encodings.

3.2.2 Software Requirements:

o Operating System: Windows/Linux/MacOS.
o Programming Language: Python 3.x. [10]
o Libraries/Frameworks:
• face_recognition for detection and recognition.
• OpenCV for image processing.
• Tkinter for building the user interface. [10]
• pandas for managing logs and attendance data.

8
 CHAPTER 4
Implementation and Result

4.1 Snap Shots of Result:

Figure 1: Snapshot of the Home Page Interface for the

Face Recognition-Based Attendance System

9
 Figure 2: Add the student details after selecting
student details from the main menu, and click save a prompt
window will show up if successful.

Figure 3: Success! Data is saved in the local database and

can be visualized in the table at right.

10
Figure 4: Camera Feed Display Before Attendance Registration

11
 Figure 5: See the saved data by importing the attendance csv

4.2 GitHub Link for Code:

https://github.com/Nandhida/Attendance-management-system-using-face-recognition.git

12
 CHAPTER 5
Discussion and Conclusion

5.1 Future Work:

Large-Scale Face Recognition for More Than 100 Faces Simultaneously

1. Efficient Preprocessing:
o Utilize multi-threading or parallel processing to handle video frames concurrently for
high throughput.
o Apply batch face encoding rather than processing faces one by one, significantly
improving computational efficiency.
2. Optimized Algorithms:
o Switch to lightweight, scalable models like MobileFaceNet or YOLOv8 for faster detection
without compromising accuracy.
o Use approximate nearest neighbor (ANN) algorithms like FAISS for matching face
encodings rapidly, ideal for large datasets.
3. Hardware Scaling:
o Deploy the system on a GPU-enabled server to handle simultaneous detections and
encodings.
o Integrate multiple cameras with distributed processing to expand coverage in larger
environments.
4. Cloud-Based Deployment:
o Store face encodings in a cloud database (e.g., Firebase, AWS DynamoDB) to
ensure scalability.
o Use cloud computing services for real-time data analysis and recognition, reducing local
computational overhead.

Enhancing Efficiency and Robustness

1. Advanced Techniques:
o Implement active learning to continually improve the recognition model by
retraining on difficult-to-classify faces.
o Integrate pose estimation to handle varying angles, enhancing recognition
accuracy under real-world conditions.
2. Dynamic Adaptation:
o Introduce adaptive frame skipping during continuous recognition, reducing
redundant computations while maintaining accuracy.
3. Error Handling:

13
 o Develop robust mechanisms to detect and flag errors, such as duplicate recognition or
lighting-induced misclassification.

User Interface (UI) Enhancements

1. Modernized Design:
o Replace Tkinter with a more flexible and visually appealing framework like PyQt5
or Flutter.
o Introduce drag-and-drop features for uploading face images and managing
data intuitively.
2. Accessibility Features:
o Add multilingual support for diverse user bases.
o Incorporate voice-guided navigation for hands-free operation.
3. Interactive Logs and Reporting:
o Allow real-time visualization of attendance trends with interactive graphs and
filters.
o Enable bulk export of logs in multiple formats (CSV, Excel).

Future Work

1. Integration with IoT Devices:

o Link with IoT-enabled attendance gates or turnstiles for seamless entry and exit
management.
o Use smart lighting or signaling systems to indicate attendance status visually in
large halls.
2. Face Recognition in Challenging Conditions:
o Deploy infrared cameras for better detection in low-light
environments.
o Develop occlusion-resilient models to recognize partially visible faces.
3. Data Privacy and Security:
o Encrypt stored face encodings using secure algorithms like AES.
o Implement GDPR-compliant policies to ensure user data safety and anonymity.

14
5.2 Conclusion:

The Face Recognition-Based Attendance Management System is a powerful solution that

leverages advanced computer vision and machine learning techniques to streamline the
traditional attendance process. This project highlights the intersection of technology and
practicality, demonstrating its potential to improve efficiency, accuracy, and ease of use in
educational institutions, workplaces, and event management scenarios.

The system's core features—real-time face detection, accurate recognition, color-coded

feedback, and robust attendance logging—are designed for simplicity and effectiveness. The
intuitive user interface, built using Tkinter, ensures accessibility for a wide range of users, from
administrators to end-users.

This project lays a strong foundation for future improvements, including:

1. Scalability: Enhancing the system to handle hundreds of faces simultaneously with

cloud-based storage and processing.
2. Advanced Recognition: Overcoming challenges like poor lighting, occlusion, and
facial variations using advanced algorithms and hardware.
3. Improved User Experience: Upgrading the interface for a more modern,
dynamic, and user-friendly experience.
4. Data Privacy and Security: Ensuring the system adheres to strict privacy
standards with robust encryption and secure storage mechanisms.

The successful implementation of this project demonstrates how AI and face recognition
technologies can automate and improve mundane tasks while paving the way for broader
applications such as security systems, personalized services, and smart environments.

By addressing limitations and expanding functionalities, this project has the potential to evolve
into a comprehensive attendance management platform with multi-domain applications,
embodying a forward-thinking approach to real-world problems.

15
 REFERENCES

[1] C. Anilkumar, B. Venkatesh, and S. Annapoorna, [1] in 2023 Second International

Conference on Augmented Intelligence and Sustainable Systems (ICAISS), IEEE, Aug.
2023, pp. 1149–1155. doi: 10.1109/ICAISS58487.2023.10250715.

[2] J. P. Jeong, M. Kim, Y. Lee, and P. Lingga, “IAAS: IoT-Based Automatic Attendance System
with Photo Face Recognition in Smart Campus,” in 2020 International Conference on
Information and Communication Technology Convergence (ICTC), IEEE, Oct. 2020, pp. 363–
366. doi: 10.1109/ICTC49870.2020.9289276.

[3] S. Kakarla, P. Gangula, M. S. Rahul, C. S. C. Singh, and T. H. Sarma, “Smart Attendance

Management System Based on Face Recognition Using CNN,” in 2020 IEEE-HYDCON, IEEE,
Sep. 2020, pp. 1–5. doi: 10.1109/HYDCON48903.2020.9242847.

[4] A. Raghuwanshi and P. D. Swami, “An automated classroom attendance system using
video based face recognition,” in 2017 2nd IEEE International Conference on Recent Trends
in Electronics, Information & Communication Technology (RTEICT), IEEE, May 2017, pp.
719–724. doi: 10.1109/RTEICT.2017.8256691.

[5] S. Sawhney, K. Kacker, S. Jain, S. N. Singh, and R. Garg, “Real-Time Smart Attendance
System using Face Recognition Techniques,” in 2019 9th International Conference on Cloud
Computing, Data Science & Engineering (Confluence), IEEE, Jan. 2019, pp. 522–525. doi:
10.1109/CONFLUENCE.2019.8776934.

[6] K. Painuly, Y. Bisht, H. Vaidya, A. Kapruwan, and R. Gupta, “Efficient Real-Time Face
Recognition-Based Attendance System with Deep Learning Algorithms,” in 2024
International Conference on Intelligent and Innovative Technologies in Computing,
Electrical and Electronics (IITCEE), IEEE, Jan. 2024, pp. 1–5. doi:
10.1109/IITCEE59897.2024.10467743.

[7] Z. Yu, Y. Qin, X. Li, C. Zhao, Z. Lei, and G. Zhao, “Deep Learning for Face Anti-Spoofing: A
Survey,” IEEE Trans Pattern Anal Mach Intell, pp. 1–22, 2022, doi:
10.1109/TPAMI.2022.3215850.

[8] K. M. M. Uddin, A. Chakraborty, Md. A. Hadi, M. A. Uddin, and S. K. Dey, “Artificial

Intelligence Based Real-Time Attendance System Using Face Recognition,” in 2021 5th
International Conference on Electrical Engineering and Information Communication
Technology (ICEEICT), IEEE, Nov. 2021, pp. 1–6. doi: 10.1109/ICEEICT53905.2021.9667836.

[9] H. Sultan, M. H. Zafar, S. Anwer, A. Waris, H. Ijaz, and M. Sarwar, “Real Time Face
Recognition Based Attendance System For University Classroom,” in 2022 2nd
International Conference on Artificial Intelligence (ICAI), IEEE, Mar. 2022, pp. 165–168. doi:
10.1109/ICAI55435.2022.9773650.

16
[10] A. M. Tripathi, A. K. Rai, and D. Pandey, “Face Recognition-Based Automated Attendance
System,” in 2024 IEEE International Conference on Computing, Power and Communication
Technologies (IC2PCT), IEEE, Feb. 2024, pp. 563–565. doi:
10.1109/IC2PCT60090.2024.10486785.

17