A

Minor Project Synopsis on

Title of Project Report

Submitted to

RAJIV GANDHI PROUDYOGIKI VISHWAVIDYALAYA, BHOPAL (M.P)

In Partial fulfillment for the award of degree of

BACHELOR OF TECHNOLOGY
IN
CSE /AIML/IOT/IT (WRITE FULL NAME OF DEPT)

By
Prachi Mavkar, 0818CS221134
Gungun Jaiswal,0818CS221079
Piyush Pushpad,0818CS221132
Mahi Changal,0818CS1107

Under the Guidance of

Mr AMIT KANUNGO
Designation
 INDORE INSTITUTE OF SCIENCE & TECHNOLOGY, INDORE
PITHAMPUR ROAD, OPPOSITE IIM, RAU, INDORE 45333, MP.
Approved by AICTE, New Delhi, affiliated to RGPV, Bhopal, Recognized by UGC under Section 2(f)

DECLARATION

We hereby declare that the work, which is being presented in this dissertation
entitled “Face recognition online attendance” in partial fulfillment of the
requirements for the award of the degree of Bachelor of Technology in
Computer science and engineering, is an authentic record of work carried out
by us.

The matter embodied in this report has not been submitted by us for the award of
any other degree.

Prachi Mavkar Gungun Jaiswal

0818CS221134 0818CS221079

Piyush Pushpad Mahi Changal

0818CS221132 0818CS221107
 INDORE INSTITUTE OF SCIENCE & TECHNOLOGY, INDORE
PITHAMPUR ROAD, OPPOSITE IIM, RAU, INDORE 45333, MP.
Approved by AICTE, New Delhi, affiliated to RGPV, Bhopal, Recognized by UGC under Section 2(f)

Department of Computer Science & Engineering

CERTIFICATE

This is to certify that the project entitled “ Face recognition online attendance”
submitted to RGPV, Bhopal (M.P.) in the Department of Computer science and
engineering by

Mr./Ms PRACHI MAVKAR Enrollment number :0818CS221134

Mr./Ms GUNGUN JAISWAL Enrollment number :0818CS221079

Mr./Ms PIYUSH PUSHPAD Enrollment number :0818CS221132

Mr./Ms MAHI CHANGAL Enrollment number :0818CS221107

in partial fulfillment of the requirement for the award of the degree of Bachelor of
Technology in COMPUTER SCIENCE AND ENGINEERING during the academic
year 2022-2026

PROJECT GUIDE PROJECT COORDINATOR

 INDORE INSTITUTE OF SCIENCE & TECHNOLOGY, INDORE
PITHAMPUR ROAD, OPPOSITE IIM, RAU, INDORE 45333, MP.
Approved by AICTE, New Delhi, affiliated to RGPV, Bhopal, Recognized by UGC under Section 2(f)

ABSTRACT

The Face Recognition Online Attendance System is an advanced technological solution designed to
automate and streamline the attendance process using facial recognition algorithms. Traditional
attendance systems, such as manual sign-ins or RFID-based systems, are prone to human errors,
time theft, and inefficiency. This project leverages machine learning techniques, specifically deep
learning-based facial recognition, to authenticate and mark attendance based on a student’s or
employee's face. The system employs high-accuracy facial recognition models to compare a
captured image of the individual with a pre-enrolled database, ensuring real-time attendance
recording.
The application is integrated into an online platform that allows teachers, administrators, or
employers to monitor attendance remotely, providing a secure and efficient way to track presence.
It eliminates the need for physical interaction with attendance systems, ensuring safety and saving
time. The system is easy to deploy, scalable, and can be integrated with existing infrastructure. The
use of this automated attendance system increases accuracy, reduces administrative workload, and
ensures transparency. Furthermore, it offers significant potential for various applications, such as
education institutions, corporate offices, and government agencies, offering a modern, reliable
solution for attendance management.
 INDORE INSTITUTE OF SCIENCE & TECHNOLOGY, INDORE
PITHAMPUR ROAD, OPPOSITE IIM, RAU, INDORE 45333, MP.
Approved by AICTE, New Delhi, affiliated to RGPV, Bhopal, Recognized by UGC under Section 2(f)

Acknowledgement

We take this opportunity to express our heartfelt gratitude to everyone who has
supported us throughout the journey of completing our minor project.

First and foremost, we would like to thank the Head of the Department, Dr
RICHA GUPTA, for their encouragement, guidance, and providing us with all
the necessary facilities to carry out this project.

We extend our sincere gratitude to our Project Coordinator, Mrs PRAVEENA

JOSHI, for their continuous support, timely advice, and valuable feedback. We
are deeply indebted to our Project Guide, MR AMIT KANUNGO, for their
expert guidance, valuable suggestions, and support throughout this project.

Thank you all for making this project a meaningful and enriching experience.

PRACHI MAVKAR GUNGUN JAISWAL

0818CS21134 0818CS221079

PIYUSH PUSHPAD MAHI CHANGAL

0818CS221132 0818CS221107
 INDORE INSTITUTE OF SCIENCE & TECHNOLOGY, INDORE
PITHAMPUR ROAD, OPPOSITE IIM, RAU, INDORE 45333, MP.
Approved by AICTE, New Delhi, affiliated to RGPV, Bhopal, Recognized by UGC under Section 2(f)

INDEX

CHAPTER PAGE NO

Abstract i
Acknowledgement ii
Table of Contents vi
1. Introduction 1
1.1 Background
1.2 Problem Statement
1.3 Purpose/Objective
1.4 Solution Approach
1.5 Scope of the Project
2. Problem Domain
2.1 Limitations of Current Systems

2.2 Need for Improvement

3. Solution Domain
3.1 Introduction
3.2 Architecture and Framework
3.3 Techniques and Tools used
4. Requirements and Analysis
4.1 Feasibility Study
4.2 Functional Requirements
4.3 Non-Functional Requirements
5. Conclusion
References
1. Introduction

In educational institutions, traditional attendance methods often involve manual

roll-calling or paper-based registers. These methods are time-consuming, prone
to errors, and susceptible to proxy attendance. This project leverages face
recognition technology to provide a modern solution for automated attendance,
aiming for efficiency and accuracy.

1.2 Problem Statement

Manual attendance tracking leads to inefficiencies and can lack security. There is a need for an
automated attendance system that ensures accurate tracking, prevents unauthorized attendance, and
reduces administrative burden.

1.3 Purpose
The purpose of this project is to develop a reliable and secure attendance system that automates the
recording process by identifying student faces. This project aims to enhance attendance
management by ensuring an error-free and secure

1.4 Solution approach

The Face Recognition Attendance System captures and stores images of students during
registration, then uses a camera to recognize registered faces and log attendance automatically. It
operates within a time-bound attendance window, during which the camera is active for capturing
attendance.

1.5 Scope of the Project

This project is designed for classrooms, allowing students and teachers to access attendance data
via a user-friendly interface. With scalability as a priority, the system has potential for future
integration with other school management tools.

2. Problem Domain

Limitations of Current Systems:

. Accuracy Issues
• False Positives/Negatives: Face recognition systems can sometimes fail to accurately
identify individuals, leading to false positives (incorrectly recognizing someone) or
false negatives (failing to recognize the right person). This is especially problematic in
systems that are not well-calibrated or have poor quality input data.
• Lighting Conditions: Poor lighting, shadows, or overly bright backgrounds can
significantly reduce the accuracy of face recognition systems. Face detection
performance tends to decrease under non-ideal lighting conditions.
• Low-Resolution Images: Low-quality or low-resolution images/videos can make it
difficult for the system to capture distinct facial features, leading to errors in
identification.
Liveness Detection Challenges
• Spoofing: Face recognition systems are vulnerable to spoofing attacks, where
individuals use photos, videos, or masks to impersonate others. This is a major
limitation in systems that do not have robust liveness detection mechanisms.
• Evasion Attacks: More sophisticated spoofing methods, such as 3D masks or deepfake
technologies, are harder to detect by traditional face recognition systems, making them
less secure.
3. Privacy and Security Concerns
• Data Security: Storing and transmitting biometric data, such as facial images, raises
concerns about data breaches and unauthorized access. If not properly secured, facial
data could be vulnerable to hacking or misuse.
• Privacy Issues: Some users may have concerns about the collection and storage of their
biometric data. Without proper consent or transparency, face recognition systems can
violate privacy rights.
• Bias and Fairness: Many face recognition systems have been criticized for being biased
towards certain demographics (e.g., race, gender, age). This is often due to biases in
 the training data, leading to unfair results for individuals outside the dominant
demographic group.
4. Environmental Factors
• Face Orientation: Current systems may struggle with detecting faces that are turned
at unusual angles or if the person is wearing accessories like hats, glasses, or masks.
This affects the system’s ability to correctly capture facial features.
• Multiple People in the Frame: When several individuals are present in the camera's
field of view, the system may fail to correctly distinguish between them, especially if
their faces overlap or if they are too far away.
5. Real-Time Processing Limitations
• Processing Speed: Real-time processing of face recognition can require significant
computational power, which may not always be available. This can lead to delays or
lag in detecting faces and marking attendance, particularly in larger environments.
• Hardware Limitations: High-performance cameras, sensors, and processing units are
required for accurate real-time face recognition. Not all institutions or organizations
have the necessary infrastructure to support these systems.
6. Scalability Issues
• Large-Scale Deployment: For large organizations, schools, or universities, scaling face
recognition attendance systems to handle thousands of users can become problematic.
Handling large databases and ensuring the system can perform at scale without
compromising accuracy or speed is a challenge.
• Database Management: Storing, managing, and updating the facial data of many users
in a secure and efficient way is a significant challenge. Ensuring that users’ data
remains consistent and updated across all devices and locations can be complex.
7. User Acceptance and Trust

2.2 Need for Improvement

While face detection for online attendance systems offers significant benefits, there is a clear
need for continuous improvement in various areas to ensure better performance, security,
and user experience. Below are some key areas where improvements are needed:
1. Improved Accuracy
 • Enhanced Recognition Algorithms: Current systems sometimes fail to accurately
recognize faces due to factors like lighting, resolution, or angle. There is a need for
more sophisticated machine learning algorithms, such as deep learning or hybrid
models, which can handle variations in lighting, orientation, and facial expressions
more effectively.
• Better Handling of Environmental Factors: The system should improve its ability to
detect faces under challenging environmental conditions, including low light, shadows,
and glare. This can be achieved by developing adaptive lighting technologies or using
infrared cameras for clearer detection in poor lighting conditions.
2. Liveness Detection and Anti-Spoofing Measures
• More Robust Anti-Spoofing Techniques: Face recognition systems should integrate
advanced liveness detection techniques to detect spoofing attempts using photos,
videos, or masks. Technologies such as 3D face scanning, depth-sensing cameras, and
motion analysis can help ensure that the person being detected is physically present.
• Continuous Monitoring: Systems can incorporate continuous biometric authentication
to monitor changes in the face during attendance to prevent spoofing attacks. AI-
powered analysis of blinking or subtle movements could improve liveness detection
accuracy.
3. Privacy and Security Enhancements
• Data Encryption and Secure Storage: Biometric data, especially facial images, needs to
be encrypted both during transmission and storage. End-to-end encryption and
decentralized storage could help improve security and ensure that personal data is not
vulnerable to breaches or misuse.
• Anonymization: To protect privacy, the system could use techniques like face
anonymization or hashing facial data before storing it, ensuring that the actual
biometric data is not directly accessible.
• User Consent and Transparency: There needs to be clear communication with users
regarding how their data is being collected, stored, and used. Transparent privacy
policies and user consent mechanisms will help users feel more comfortable using face
recognition systems.
4. Reducing Bias and Improving Fairness
• Diverse Training Datasets: One of the major criticisms of current face recognition
systems is that they tend to be biased towards certain demographics (e.g., race, gender,
 or age). Improving the diversity of training datasets will help the system to more
accurately recognize faces across different ethnicities, genders, and ages, ensuring
fairness and reducing errors.
• Bias Testing and Auditing: Continuous auditing of the system's performance to
identify and correct biases is critical. Developers should perform regular bias testing
to ensure the system works equally well for all user groups.
5. Improved User Experience
• User-Friendly Interfaces: The user interface (UI) should be intuitive and simple,
making the system easy for non-technical users to interact with. This includes
improving the camera setup to ensure that users are correctly aligned with the camera
and receive feedback if adjustments are necessary.
• Multi-Factor Authentication (MFA): To improve accuracy and security, face
recognition could be integrated with multi-factor authentication methods such as
PINs, QR codes, or fingerprints. This would provide an additional layer of security
and mitigate issues where face recognition alone may not be reliable.
6. Scalability and Real-Time Processing
• Handling Large-Scale Deployments: For large organizations, improving scalability is
critical. Systems should be designed to handle a large number of users without
performance degradation. This may require the use of cloud-based solutions, edge
computing, or distributed databases to ensure fast and efficient processing of large
datasets.
• Faster Processing Times: Real-time face recognition can be resource-intensive.
Improvements in the system's speed, such as through the use of faster GPUs or
optimized software frameworks, can ensure that large-scale attendance systems
operate without delays or lag, providing instant results.
7. Adaptability to Changing Conditions
• Dynamic Face Recognition Models: Face recognition systems should be able to adapt
over time to accommodate changes in a person's appearance, such as aging, changes in
hairstyle, or the use of accessories like glasses or masks. Implementing dynamic
models that can update or retrain automatically as users' facial features evolve will
improve long-term accuracy.
• Facial Variation Handling: The system could use more advanced techniques such as
3D face recognition or thermal imaging to capture more details and accurately identify
 individuals, even if their appearance changes due to lighting, health conditions, or
other factors.

3.Solution Domain

3.1 Introduction
The solution domain of face recognition online attendance refers to the area or field of
technologies, methodologies, and systems that enable attendance tracking through face
recognition in an online or digital context. This solution combines multiple advanced
technologies such as computer vision, machine learning, deep learning, and cloud
computing to automate attendance processes, ensuring accuracy, convenience, and
efficiency.

3.2 Architecture and Framework

The architecture and framework for a Face Recognition Online Attendance

System typically involve various layers of components that work together to perform
automated face recognition, store data, and manage attendance records. Below is a
breakdown of the architecture and framework used to implement such a system:
1. System Architecture
The system architecture can be divided into several major components:
1.1 User Layer (Client-Side)
• Users: Students, employees, or any individuals whose attendance is being
tracked.
• Devices: Laptops, desktops, tablets, or mobile phones with cameras that capture
images or videos for face recognition.
• UI/UX:
 o Web/Mobile Interface: An application or website where users can
check their attendance status, login, or register.
o Real-time Feedback: The system provides immediate feedback when a
user’s face is recognized (attendance is marked or denied).
1.2 Camera and Face Recognition Module (Capture and Processing Layer)
• Cameras: High-definition cameras that capture images/videos of users (could be
webcam or external cameras).
• Face Detection Algorithm: Detects human faces in the captured images or
video frames. Popular libraries like OpenCV or Dlib are used for face
detection.
• Face Recognition Algorithm: Compares the detected face with stored facial
data to identify the person. Libraries like TensorFlow or FaceNet are often
used to implement deep learning models for face recognition.
1.3 Server Layer (Backend and Business Logic Layer)
• Face Recognition Service:
o Once a face is captured, the system compares it against a pre-stored
facial database.
o It uses machine learning models (such as Deep Learning models,
CNNs, or Siamese Networks) to perform accurate face matching.
• Attendance Management System (AMS):
o Database: Stores the attendance data (user ID, time, date, etc.) and
biometric data (face embeddings).
o API Server: Exposes RESTful APIs to integrate with the front end
(web/mobile apps) and other systems like LMS or ERP for syncing
attendance data.
• Authentication & Authorization: Ensures that only authorized users (students,
employees, or admins) can access the system and mark attendance.
1.4 Database Layer (Data Storage)
 • Relational Database (SQL): For storing user details, attendance logs,
timestamps, and any other structured data (e.g., MySQL, PostgreSQL).
• NoSQL Database: For storing large amounts of unstructured data like facial
recognition embeddings (e.g., MongoDB, Firebase).
• Data Encryption: Sensitive data (especially biometric data) is encrypted for
security and privacy.
1.5 Cloud Layer (Optional)
• Cloud Storage: The database and user data can be stored in the cloud (AWS,
Google Cloud, Azure) for scalability and redundancy.
• Cloud-based Face Recognition Models: Some services (e.g., AWS
Rekognition, Google Vision API) offer face recognition APIs that can be used
instead of deploying custom-trained models.
• WebSocket for Real-Time Data: For real-time interaction (e.g., marking
attendance or sending notifications), a WebSocket or push notifications are
implemented.
1.6 Admin Layer (Management and Analytics Layer)
• Admin Dashboard: For managing attendance records, users, and system
settings.
o Attendance Reports: Detailed reports of student or employee
attendance history.
o System Monitoring: Monitoring the system's health, usage, and errors.
• Notifications: Admins or users may receive alerts in case of issues, like failed
attendance capture or multiple attempts to mark attendance.
Frameworks & Tools Used
2.1 Frontend Frameworks
• Web Frameworks:
o React or Vue.js for dynamic, real-time user interfaces.
o HTML/CSS/JavaScript for rendering the user interface.
 o Bootstrap or Material UI for responsive design.
• Mobile Frameworks (if a mobile app is used):
o Flutter or React Native for cross-platform app development.
2.2 Backend Frameworks
• Server Framework:
o Node.js with Express.js for creating the server and handling HTTP
requests.
o Python with Flask or Django (for Python-based face recognition
systems).
• Face Recognition Libraries/Tools:
o OpenCV: Open-source computer vision library for face detection and
image processing.
o Dlib: A toolkit for machine learning, offering facial landmark detection
and face recognition.
o TensorFlow/Keras: For training and deploying deep learning models
for face recognition.
o FaceNet: A popular model for face recognition and embedding
extraction.
o MTCNN or Haar Cascade: Popular algorithms for face detection.
2.3 Machine Learning Libraries
• Deep Learning Models: Use pre-trained models such as VGG-Face, FaceNet,
or ResNet for recognizing faces.
• Model Deployment:
o TensorFlow Serving or Flask to serve the trained face recognition
models for predictions.
2.4 Databases
• MySQL/PostgreSQL: To store structured data such as user profiles and
attendance logs.
• MongoDB: If you prefer a NoSQL database to store unstructured data like face
embeddings.
• Firebase: For real-time synchronization and cloud storage of data.
2.5 Cloud Services (Optional)
• AWS Rekognition: For facial analysis and recognition.
• Google Cloud Vision: Another option for face recognition using cloud services.
• Microsoft Azure Face API: A cloud-based service for face detection and
recognition.

3. System Workflow
1. User Registration:
o The user (student/employee) registers their face in the system, where
multiple images of their face are captured from different angles for
better recognition accuracy.
o These images are processed, and facial embeddings (a unique numerical
representation of the face) are generated and stored in the database.
2. Face Capture and Attendance Marking:
o When the user appears in front of the camera (either via a mobile or
desktop camera), the system detects their face and captures it.
o The captured face image is processed and matched against the database.
o If the system finds a match, the user’s attendance is marked as present.
If no match is found, the system may notify the user or mark the
attendance as failed.
3. Data Storage and Reporting:
o Attendance data is saved in the database, where admins can access
reports.
o Analytics tools can be applied to generate insights into attendance
patterns.
 4. Security & Privacy Considerations
• Data Encryption: Both face data and personal information should be encrypted
using advanced encryption algorithms (e.g., AES-256).
• GDPR/Legal Compliance: Ensure that the system complies with local data
protection laws, such as GDPR, when handling biometric data.
• Access Control: Implement strong authentication and role-based access controls
for admins and users.

3.3 Techniques and Tools used

4. Technology Stack
4.1 Front-End Technology
• HTML, CSS, JavaScript: Provides a simple, user-friendly interface for interaction
with the system.

4.2 Back-End Technology

• Django: Serves as the backend framework, managing requests and processing data
securely.
• FaceNet: Used as the face recognition model, enabling accurate identification of faces.

4.3 Database

• MongoDB: Stores user data, face images, and attendance records, chosen for its
scalability and flexible data management.

4.4 APIs and Other Tools

• OpenCV: Assists in image processing and face detection.
• AWS EC2: Provides a reliable deployment environment for cloud hosting and scalable
access.
• Docker: Ensures consistent development and production environments.

4. Requirements and Analysis

3.1 Problem Definition
To provide an effective solution for automated attendance, the system needs to recognize
and record student attendance through face recognition. It should reduce errors and
administrative tasks while enhancing data security and accessibility.
3.2 Feasibility Study
• Technical Feasibility: The project is technically feasible using existing tools and
frameworks such as Django for the backend, FaceNet for face recognition, and MongoDB for
storage.
• Economic Feasibility: Initial development costs are low, as open-source tools and cloud
deployment (e.g., AWS EC2) are utilized.
• Operational Feasibility: The system is straightforward and fits seamlessly into daily
operations. Teachers and students can easily use the system with minimal training.

3.3 Functional Requirements

• User Registration: Students and teachers register through the system, with face images
captured and stored.

• Real-time Face Recognition: The system identifies faces in real-time within a specified
window after the class begins.
• Attendance Reports: Attendance data can be accessed by teachers for tracking and
record-keeping.

3.4 Non-Functional Requirements

• Performance: Must handle multiple users and process images quickly.
• Scalability: Capable of supporting additional students and expanding functionality.
• Usability: Designed to be user-friendly and intuitive for both students and teachers.
• Reliability and Availability: Ensures high uptime, especially during class hours.
• Security: Access-controlled, with encrypted data storage to protect sensitive
information.

4. Conclusion
The Face Recognition Attendance System project is a comprehensive solution for
modernizing attendance processes. It provides automated attendance tracking with real-
time face recognition, ensuring secure, reliable, and efficient data handling. Future
developments could transform it into an ERP tool, offering a wide range of management
solutions for educational institutions.

5 .REFRENCES

• FaceNet documentation and GitHub repository

• Django official documentation
• MongoDB manual for database management
• OpenCV library documentation
• AWS documentation for EC2 deployment