Smart Attendance Management System

using Computer Vision

A Project Report
Submitted
In Partial Fulfillment of the
Requirements For the Degree of

Bachelor of Technology (B.Tech)

in
Computer Science & Engineering
by

Peehoo Shrotriya 2101920100197

Nitin Kumar Tiwari 2101920100192
Priyansh Saxena 2101920100214

Under the Supervision of

Dr. Nitin Tyagi
Professor

G. L. BAJAJ INSTITUTE OF TECHNOLOGY & MANAGEMENT

GREATER NOIDA

DR. A. P. J. ABDUL KALAM TECHNICAL UNIVERSITY,

UTTAR PRADESH, LUCKNOW
2024-2025
 Declaration

We hereby declare that the project work presented in this report entitled “Smart Attendance
Management System using Computer Vision”, in partial fulfillment of the requirement for
the award of the degree of Bachelor of Technology in Computer Science & Engineering,
submitted to Dr. A.P.J. Abdul Kalam Technical University, Uttar Pradesh , Lucknow is based
on our own work carried out at Department of Computer Science & Engineering, G.L. Bajaj
Institute of Technology & Management, Greater Noida. The work contained in the report is
true and original to the best of our knowledge and project work reported in this report has not
been submitted by us for award of any other degree or diploma.

Signature:
Name: Peehoo Shrotriya

Roll No: 2101920100197

Signature:
Name: Nitin Kumar Tiwari

Roll No: 2101920100192

Signature:
Name: Priyansh Saxena

Roll No: 2101920100214

Date:
Place: Greater Noida

ii
 Certificate

This is to certify that the Project report entitled “Smart Attendance Management System

using Computer Vision” done by Peehoo Shrotriya (2101920100197), Nitin Kumar

Tiwari (2101920100192), and Priyansh Saxena (2101920100214) is an original work

carried out by them in Department of Computer Science & Engineering, G.L. Bajaj Institute

of Technology & Management, Greater Noida under my guidance. The matter embodied in

this project work has not been submitted earlier for the award of any degree or diploma to the

best of my knowledge and belief.

Date:

Dr. Nitin Tyagi Dr. Sansar Singh Chauhan

Signature of the Supervisor Head of the Department

iii
 Acknowledgement

The merciful guidance bestowed to us by the almighty made us stick out this project to a
successful end. We humbly pray with sincere heart for his guidance to continue forever.

We pay thanks to our project guide Dr. Nitin Tyagi who has given guidance and light to us
during this project. His versatile knowledge has helped us in the critical times during the span
of this project.

We pay special thanks to our Head of Department Dr. Sansar Singh Chauhan who has been
always present as a support and help us in all possible way during this project.

We also take this opportunity to express our gratitude to all those people who have been directly
and indirectly with us during the completion of the project.

We want to thanks our friends who have always encouraged us during this project.

At the last but not least thanks to all the faculty of CSE department who provided valuable
suggestions during the period of project.

iv
 Abstract

Smart Attendance Management System is a very innovative solution for the automation of

attendance within educational institutions. The primary intention is to reduce manual

intervention and improve accuracy and efficiency in tracking student and teacher attendance.

The system allows teachers to mark attendance digitally, automatically generate attendance

records based on course schedules, and even provide capabilities to edit or update the records

as needed. The system also has full control by the admins as they manage student registrations,

teachers' assignment to courses, timetables, and attendance data.

It also supports easy attendance export into formats like Excel and PDF to support easy

reporting and analysis. The system has a student management module in which admins can

manage the data of students, course offerings, timetables, and teacher assignments. Teachers

interact with the system to monitor the attendance of students, features like start/stop

attendance, and manual adjustments to any discrepancy.

This system removes the need for traditional paper-based attendance methods, offers real-time

tracking, and helps generate accurate attendance reports, thereby reducing administrative

overhead. Integration of modern technologies ensures secure and scalable operations for

educational institutions, providing a user-friendly platform that can be accessed by both faculty

and administrators. SAMS is designed to improve overall institutional productivity, reduce

errors, and ensure better attendance tracking for both academic and administrative purposes.

v
 TABLE OF CONTENT

Declaration …………………………………………………………………………… (ii)

Certificate …………………………………………………………………………… (iii)
Acknowledgement ………………………………………………………………………. (iv)
Abstract ………………………………………………………………………….. (v)
Table of Content ……………………………………………………………………….. (vi)
List of Figures …………………………………………………………………………... (viii)
List of Abbreviations …………………………………………………………………... (ix)
Chapter 1. Introduction …………………………………………………………….. 1
1.1 Preliminaries……………………………………………………………... 1
1.2 Problem Analysis………………………………………………………… 2
1.3 Motivation………………………………………………………………... 3
1.4 Objectives………………………………………………………………… 4
Chapter 2. Literature Survey……………………………………………………….. 5
2.1 Introduction ……………………………………………………………... 5
2.2 Existing Systems…………………………………………………………. 7
2.3 Strengths of the Project………………………………………………...… 9

Chapter 3. Proposed Methodology…………………………………………………. 11

3.1 Problem Formulation ………………………………………………….… 11
3.2 System Flow……………………………………………………………… 12
3.3 Flow Chart……………………………………………………………….. 14
3.4 Use Case Diagram………………………………………………………... 15
3.5 Entity Relationship Diagram .…………………………………………… 20
3.6 System Flow Components……………………………………………….. 26
3.7 Proposed Work………………………………………….……………….. 27

Chapter 4. Implementation ………………………………………………………… 30

4.1 Introduction ……………………………………………………………… 30
4.2 Implementation Strategy…………………………………………………. 30
4.3 Tools/Hardware/Software Requirements.………………………………... 40
4.4 Expected Outcome ………………………………………………………. 42

Chapter 5. Result & Discussion.……………………………………..……………...

44
5.1 Result…………………………………………………………………….. 44
5.2 Discussion………………………………………………………………... 48
5.3 Integration Challenges & Mitigation Strategies…………………………. 49

vi
Chapter 6. Conclusion & Future Scope……………………………………………. 52
6.1 Conclusion……………………………………………………………….. 52
6.2 Future Scope…………………………………………………………….. 53
References………………………………………………………………………………… 57

vii
 LIST OF FIGURES

Figure No. Description Page No.

Figure 3.1 System Design Flow 12
Figure 3.2 Process Flowchart 14
Figure 3.3 Use Case Diagram for SAMS 16
Figure 3.4 SAMS Database Structure 22
Figure 4.1 Smart Attendance Management System 31
Figure 4.2 Load Students 35
Figure 4.3 Encodings 36
Figure 4.4 Valid Student 37
Figure 4.5 Attendance 38
Figure 4.6 Mark Automatic Attendance 39
Figure 5.1 Landing Page 44
Figure 5.2 Registration Form 1 45
Figure 5.3 Registration Form 2 45
Figure 5.4 Dashboard 46
Figure 5.5 Report Generate View 47
Figure 5.6 Report PDF Form 47

viii
 LIST OF ABBREVIATIONS

Abbreviations Full Form Page No.

SAMS Smart Attendance Management System 1
RFID Radio Frequency Identification 2
AI Artificial Intelligence 5
ER Entity Relationship 20
CSV Comma Separated Value 34
RGB Red, Green, and Blue 36
IP Internet Protocol 42
LMS Learning Management Systems 55
SIS Student Information Systems 55

ix
x
 Chapter 1
Introduction

1.1 Preliminaries

Smart Attendance Management System, or SAMS, is the next generation of technological

solution in educational institutions. It can automatically track the attendance with the help
of the latest computer vision and machine learning technologies to get the accurate records
of student attendance in real time. Traditional methods of attending class usually occupy a
lot of class time and result in the possibility of having wrong marks against a student's
attendance or even proxy attendance wherein a student puts marks that another student put
against his when he did not attend classes in Raj et al. [1].

SAMS removes all of these problems using facial recognition technology, a very advanced
biometric identification technology, allowing the system to uniquely identify the students
by their facial features. The system uses the Python face_recognition library, which is an
advanced tool for machine learning, allowing efficient face detection and recognition. This
technology uses algorithms trained on vast datasets of human faces, enabling it to identify
individuals with a high degree of accuracy.

Once the student enters the classroom, real-time face detection is used to identify him by
crossing him with a pre-created database of all enrolled students. On successful match, the
attendance of the student is automatically marked; the process is seamless without human
intervention. The system may function using several cameras, thus recording all the
students from different angles and possibly crowded and dynamic classrooms.

1
It will not leave behind some of the administrative roles to be fulfilled; hence there will be
adequate time for students as well as their lecturers to receive proper attention and raise
their probabilities to possess up-to-date attendance records.

1.2 Problem Analysis

Attending school has for a long time been an everyday resource-intensive activity in
learning institutions that is replete with inefficiencies. Dassanayake et al. [2] discussed that
traditionally teachers have taken rolls by calling out the names of students or by having the
students sign attendance sheets by hand. These methods are both slow and error-prone.
Typical manual mistakes in attendance usually result in class records becoming wrong. In
the course of such activities, a student may be recorded absent or present when it is actually
the opposite. Another critical problem related to the system of manual attendance is proxy
attendance. In this, a student marks his peer present without attending the class. In huge
classes, where the lecturer cannot check up on every student, it is easier to use this proxy
system. Thus, proxy attendance undermines the academic assessment and record keeping
about students' attendance. This will eventually affect the general academic performance
of a student due to unaddressed absence.

While some institutions try to automate this process with technologies such as Radio
Frequency Identification or RFID cards or biometric fingerprint scanners, each of these has
its own limitations.

The RFID system consists of individual identifiers on student cards. The cards have to be
scanned for attendance. The RFID identification is easily interrupted by impediments,
hence requiring students to carry identification cards in a manner that the Zhao et al. [3]
explained. That is why the limitations is that students lose or forget their cards or get
scanned for them, which can get the entire system into chaos.

2
The RFID-based technologies have weaknesses that, if exploited, can breach the
authentication systems on university campuses. Attacks such as RFID skimming and tag
killing are simple to conduct and can have severe effects on the victims reported by Pereira
et al. [4]

Biometric fingerprint recognition has its own disadvantages, such as system accuracy
problems; for example, it may identify a wrong person due to dirt or worn-out fingerprints.
It raises privacy concerns over the collection and storage of sensitive biometric data. The
system fails to work efficiently when a student has physically altered fingerprints either
due to injury or for any other reason, which might result in false negatives.

1.3 Motivation

The motivation behind the development of the SAMS is the need to offer a more efficient,
accurate, and secure solution to the challenges posed by traditional attendance management
systems. Several key factors drive the necessity for such a system:

1. Time Effectiveness: Manual attendance marking occupies a lot of class time that
could be more effectively utilized for instruction or activities. SAMS saves this
time as it automatically marks attendance in almost no time. Once a student enters
the classroom, the system captures his face and updates the attendance records, and
the time spent on administrative tasks is kept to a minimum.

2. Accuracy and Accountability: One of the major reasons is to make sure that
attendance records are accurate and dependable. Facial recognition eliminates
human error and ensures there will not be discrepancies. It also helps in avoiding
proxy attendance where students sitting outside are marked as present inside,
increasing academic integrity and accountability.

3
It is this push towards open, efficient systems whose thrust is towards more transparent,
timely, and accurate attendance data that motivates the project. The administration
overheads in an institution become less necessary when SAMS improves its productivity
while keeping attendance data in such a manner that is resistant to human error about that.

1.4 Objectives

The SAMS will eliminate inefficiencies in traditional processes of tracking attendance

because it provides automated and secure features. There are two primary goals for this
system.

Facial recognition can automate attendance: The whole purpose of the system is to
automatically make the attendance through facial recognition. In this traditional system,
calling roll calls is substituted by a system where the face of the student is scanned and
automatically identified by the system. Once it is identified in real time, the system updates
the attendance record. Neither the teacher nor the student has anything to do. The system
ensures that only registered students are marked present and eliminates human errors or
discrepancies in attendance data.

Human error removal and saving time: A traditional attendance system has human error on
a high level, such as marking attendance for the student present or absent due to the
negligence of the human element. SAMS eliminates this because it allows accurate
marking of attendance. It saves the time taken up in taking the attendance manually. The
instructor thus has more time to teach and relate to students than doing administrative work.
Since the recognition is in real time, they have by default marked them present as there was
no delay between detection.

4
 Chapter 2
Literature Survey

2.1 Introduction

Facial recognition technology is one of the significant innovations developed in recent

years with Artificial Intelligence (AI), computer vision, and machine learning.
It is highly popular across sectors requiring efficient and accurate identification systems,
such as schools, offices, and colleges & management. Its ability to provide non-intrusive,
automated, and real-time identification solutions has made it particularly appealing in
education for managing student attendance, a historically time-intensive and error-prone
task explained by Duwal et al. [5].

According to Perwej et al.[6] that facial recognition reduces human errors, enhances
operational efficiency, and automates attendance tracking processes. Facial recognition,
unlike traditional systems that depend on manual or physical identification methods,
employs biometric identifiers—facial features—to uniquely and accurately identify
individuals. This will, therefore, make the attendance system more reliable and secure.

Real-time facial recognition systems effectively scan students' faces as they enter
classrooms, marking their attendance automatically and reducing the administrative burden
on educators.

5
Machine Learning and Real-Time Capabilities:
Facial recognition systems rely on complex machine learning algorithms, which learn and
improve accuracy as it analyzes large datasets. The algorithms process key facial
landmarks and match them to the stored data, allowing identification to be reliable have
been discussed by Yao [8]. By computer vision techniques, systems can work in real time,
making them highly viable for good-scale deployments within institutions.

Advantages Over Traditional Methods:

It is not only easier to take attendance but also ensures accountability and safety. Manual
roll calls and other methods take up more time and often include proxy attendance
mentioned by Dassanayake et al. [2]. Biometric-based systems based on fingerprint
recognition or RFID cards demand direct physical contact, which becomes inconvenient
and hygiene-bred during pandemics. This contactless and automatic mode of facial
recognition overcomes the above-mentioned constraints and has been recognized
extensively as a cost-effective, efficient, and tamper-proof alternative.

Educators' and Administrators' Perspectives:

Administrators and educators are now increasingly adopting facial recognition systems to
improve accuracy, save time, and reduce administrative overheads. Several studies
emphasize that these systems create transparent and secure attendance
records, thus fostering trust among students, parents, and institutions. Facial recognition
systems are in line with the goals of modern educational environments because they
promote efficiency, accuracy, and data-driven decision-making. This reviews the existing
attendance systems, their drawbacks, and the change that facial recognition technology can
bring to educational institutions.

6
2.2 Existing Systems

2.2.1 Manual Systems: Inefficient and Error-Prone

Manual attendance systems have been in place for years, with methods such as roll
calls or signature-based tracking. These systems, however, have several
disadvantages. According to such systems are time-consuming and error-prone,
especially for large classrooms or institutions.

Time-Consuming:
Manual roll calls take a lot of time, especially in large class. The same time
could have been utilized for teaching and discussions.

Human Error:
The manual attendance recording has an error in it. When marking attendance
manually, some students are marked present when they are absent, and vice versa.
These inaccuracies can affect performance tracking and student records.

Proxy Attendance:
One of the major disadvantages of manual systems is that proxy attendance, where
students mark the attendance of their friends is possible. This makes attendance
records unreliable and can often lead to academic dishonesty.

2.2.2 RFID-Based Systems: High Investment and Vulnerable to Abuse

The RFID-based systems were used to overcome some of the disadvantages of

manual systems. However, they have a set of disadvantages of their own.

7
High Initial Setup Cost:
Setting up RFID systems involves high infrastructure investment with RFID cards
and specialized scanners. Low-budget institutions can find these systems too costly.

Overhead for Maintenance:

RFID cards are prone to damage and loss and need to be replaced
frequently. Scanners can also be recalibrated frequently and require maintenance
to maintain their effectiveness

Vulnerability to Abuse:
Proxy attendance is still possible with RFID systems if students share their cards with
peers. This limits the system’s ability to provide secure and reliable attendance
tracking.

Privacy Concerns:
RFID systems collect data about students’ movements, raising privacy concerns
about how this data is stored and used.

2.2.3 Biometric Systems: Issues of Privacy and Maintenance

Biometric systems, like fingerprint-based or iris scan-based systems,

are relatively more secure than RFID systems. Still,
they do not come without their challenges.

Privacy Issues:
Biometric data is sensitive and must be kept secure from misuse. Institutions need to
ensure compliance with privacy regulations to safeguard the personal information of
students.

8
 Maintenance and Accuracy:
Fingerprint scanners wear out easily and environmental conditions such as dust or
moisture affect their accuracy. Maintenance is essential but adds to operational costs.

High Costs:
The specialized hardware required for biometric systems, such as scanners, is
expensive to install and maintain, which makes them less practical for large
institutions.

2.3 Strengths of the Project

The facial recognition-based Smart Attendance Management System offers several

singular benefits over other systems. The following benefits make the system a good
alternative for academic institutes in streamlining procedures relating to attendance
management:

• Accuracy for Attendance: Facial recognition technology overwrites the human

errors usually accompanying the manual attendance. Students are marked present
or absent correctly according to the facts in the classroom presence of students. The
facial recognition system can detect students instantly and accurately, reducing
chances of marking a student wrongfully absent or present. Also, the system cuts
down the possibility of proxy attendance since only the enrolled student can be
recognized. Unlike RFID or biometric systems, facial recognition requires no
physical interaction, making it faster and more hygienic.

• Reduced Administrative Burden: SAMS completes the entire process of

attendance; thus, it releases time for educators to focus more on teaching and
interacting with students. As noted, the instructors do not need to use their time in
roll calls and paper attendance sheets. This system records the attendance instantly

9
 when the student enters the classroom, and that way both educators and
administrators enjoy smooth workflow. It cuts administrative overhead, freeing the
resources of educational institutions for other productive activities such as
improved student outcomes or increased classroom engagement.

• Real-Time Updates and Reporting: SAMS offers real-time updates on attendance

data. Educators and administrators can view attendance records at any given point
in time, which enhances more dynamic reporting, for instance, detecting patterns
of students' attendance or absenteeism. The reports offered detail attendance
patterns over time. These allowed administrators to trace the extent of involvement
by the students and also pick issues that needed correction ahead of time. It helped
better to communicate with the students and their parents about attendance, always
up-to-date.

10
 Chapter 3
Proposed Methodology

3.1 Problem Formulation

In comparison to such systems, traditional attendance management is erroneous and prone

to fraudulent proxy attendance. SAMS has been designed to alleviate such problems by
using facial recognition technology in real-time for the automation of the process of
attendance tracking. Error-prone methods like manual attendance often mark the presence
of students wrongly or may omit the marking of attendance in their name due to
carelessness or failure to put signature marks. Secondly, it is quite impossible for the
teachers handling large classes to monitor the attendance of every learner regularly, which
causes proxy attendance.

The facial recognition technology will solve the above issues by ensuring the process is
safe, accurate, and automated without human intervention concerning the student's
attendance. With this face-detecting, it does so through real-time face detection and
recognition whereby it ensures that only students in the class are counted while thus
preventing fraudulent attendance cases that could arise from human mistakes. Besides that,
it reduces administrative workload, because the teacher will have a minimum opportunity
for names roll calls or handling attendance sheet in relation to teaching.

Utilizing computer vision and machine learning algorithms, SAMS will be able to offer
real-time attendance tracking solutions which will be very efficient in this regard. The same
system will recognize the incoming students in the class room based on their facial features,
which will then update attendance records as current at any given time. It negates all sorts

11
of discrepancies altogether and thus removes the flaws found in traditional systems as it is
an easy-to-deploy, scalable, and tamper-proof solution.

3.2 System Flow

Figure 3.1: System Design Flow

Figure 3.1 points out the automated attendance marking process within the Student SAMS.
It focuses on the features related to teacher interaction, attendance processing, and
reporting. The following is the explanation of each phase in the system:

Teacher Login

This involves logging into the system through credentials by the teacher. This login is role-
based and limits operations to those allowed users. The authentication of the system is
further ensured by checking fields, including username, password, and account flags such
as is_superuser, is_staff, and is_active. All these credentials are maintained in the Users
entity within the database for secure access to the attendance management system.

12
Take Attendance

Following a successful login, the teacher initiates the attendance-taking process. This
process consists of capturing the real-time attendance of students through a camera or video
feed, and this becomes an input that further gets processed. Then, the captured images
become linked to records in the Student table, containing data such as roll_no, full_name,
section, and image. This input stage ensures that the entire attendance process starts with
a correct input stage, and the system captures images that will be verified.

Inject into Model

Then the collected student images are fed into a face_recogniton library to process. The
face recognition algorithm implemented here matches the images captured against the pre-
stored images of the face in the database. Thus, image field from Student entity forms a
reference for comparison for the system and verifies the identity of all students.

Attendance Marked

Once the system has processed the input images, the attendance for students whose faces
it successfully recognized and who all belong in that class is automatically marked. The
attendance records are maintained in a specific attendance log file, with basic fields:
roll_no → Student ID
time_stamp → Exact time the attendance was recorded.

Report Generation

This ensures the automatic marking of attendance by avoiding human errors and generating
consistent and reliable attendance reports.

After marking attendance, the system allows the teacher to generate attendance reports.
The report is processed and exported, often into formats like Excel sheets for further use.
The report would has

13
Student details → roll_no, full_name
Attendance status → Present/Absent

3.3 Flow Chart

Figure 3.2: Process Flowchart

14
The process begins with Data Collection and Face Tokenization as shown in figure 3.2,
where facial images of students are collected during their registration phase. These facial
images are used to generate unique facial embeddings/tokens, ensuring a distinct identity
for each student.

Next step ensures the proper enrollment of students into the system under the Student
Enrollment phase.

Once enrolled, the system proceeds to Real-Time Attendance Marking using classroom
cameras that capture live video feeds. These feeds are processed for face detection and
recognition, where detected faces are compared with stored facial tokens using OpenCV.
If a match is found, the attendance is automatically marked in the Attendance Marking
database. Simultaneously, the real-time video feed is used to streamline attendance
monitoring.

The system further provides additional features like Automated Reporting, where
attendance data is automatically processed for reporting purposes. The attendance data can
then be exported through the Real-Time Data Export phase into formats like Excel or PDF
for offline access or reporting needs.

3.4 Use Case Diagram

The Use Case Diagram as shown in figure 3.3 described the functionalities of SAMS and
highlighted the roles the main actors:

1. Faculty

2. Admin

3. Student

15
Each actor interacts with the system to perform specific tasks based on their role and
privileges.

Figure 3.3: Use Case Diagram for SAMS

16
Faculties are responsible for managing the attendance process and may access key features
such as Login, starting and stopping attendance, generating attendance reports, editing
attendance, and exporting the final attendance data for further analysis. These
functionalities ensure that faculty members can efficiently manage real-time attendance
monitoring with accurate records.

The following use cases are available to Faculty:

• Login:
Faculty members must log in to the system to access its functionalities.
Authentication ensures only authorized personnel can mark or manage attendance.

• Start Attendance:
Faculty can initiate attendance marking for a specific course/session. This feature
likely triggers the system to recognize students using real-time video feed and facial
recognition.

• Stop Attendance:
Faculty stops the attendance marking process once the session ends.

• Generate Attendance:
Faculty can generate attendance records for a particular session, class, or date. This
includes the list of present and absent students.

• Edit Attendance:
In case of errors or special circumstances, Faculty can manually edit attendance
records to ensure accuracy.

• Export Attendance:
Faculty can export attendance data in formats such as Excel or PDF for
administrative or reporting purposes.

17
The Admin actor manages the overall system ensuring smooth operations and
configurations. The Admin can perform tasks such as Teacher Registration, Timetable
Management, and Student Management. Additionally, the Admin can Map Teachers to
Courses, which ensures accurate course assignments, and manages both Departments and
Courses within the system.

The admin performs the following use cases:

1. Login:
Admin users authenticate themselves to gain access to administrative features of
the system.

2. Teacher Registration:
Admin can register new teachers (Faculty) in the system, assigning them credentials
and linking them to specific departments or courses.

3. Time Table Management:

Admin manages and updates the timetable for courses, ensuring classes are
scheduled appropriately.

4. Student Management:
Admin handles the registration and management of students, including updating
profiles, managing courses, and verifying student details.

5. Map Teacher to Courses:

Admin links Faculty members to specific courses, ensuring they have access to
attendance functionalities for assigned subjects.

6. Department Management:
Admin manages the organization of departments in the institution, including
adding, editing, or deleting departments.

18
 7. Course Management:
Admin creates and updates course information, assigning them to specific teachers
and departments.

The Student actor has only one interaction with the system, which is in the form of the
Registration process. This would ensure that all students are correctly registered into the
system so that their attendance can be tracked.

1. Registration:
Students provide their personal information, facial data, and other relevant details
during the registration process. This data is stored in the system for future
attendance marking using facial recognition.

In summary, the system empowers Faculty to effectively track attendance, Admin to

control and manage system-wide operations and configurations, and Students to
automatically register themselves for attendance tracking.

Interactions Between Actors and Use Cases

• Faculty interacts with the system to manage attendance-related tasks, ensuring

accurate and real-time attendance records.

• Admin acts as the central manager, responsible for system setup, configuration,
and user management.

• Students only interact with the system during the initial registration phase, where
their data is captured and stored for attendance automation.

19
3.5 Entity Relationship Diagram

SAMS is an all-embracing database system well designed in order to fulfill the variable
needs of educational institutions. With its structured, highly effective way of handling
information dealing with teachers, students, class schedules, courses, and departments,
the student and academic management system streamlines the flow of various academic
functions across the boards. Its tables are also well-organized, creating foreign key
relationships that work to ensure that data exchange is smooth.

Entity Relationship (ER) Diagram as shown in figure 3.4 holds class timetables,
teachers, student records, course offers, and departmental structures within one
centralized repository of information. In the architecture, it supports integrity and
consistency in data. The operations such as marking attendance, managing timetables,
and student enrollments are streamlined. All the tables in the SAMS database have
significant contributions, and the interconnections between the tables allow the entire
system to work in coherence.

1. Timetables Table

It is an essential part of the system since it handles class schedules and makes sure that
teachers, students, and administrators are updated with regard to class information.

• Table Structure and Fields:

▪ The "id" field uniquely identifies each timetable entry.
▪ The "year" field specifies the academic year for which the schedule applies.
▪ The "section" field notes the particular course section the schedule falls
under.
▪ The "room" field denotes the physical classroom where the class will be
conducted.
▪ The "day_of_week" field notes the day of the week (such as Monday,
Tuesday) for the class.

20
 ▪ The "start_time" and "end_time" fields record the timing of the class
session to avoid scheduling conflicts.
• Foreign Key Relationships:
▪ The "teacher_id" links to the teachers table, establishing the identity of
which teacher a particular class is associated with.
▪ The "department_id" cross-links to the departments table, identifying to
which department the class falls under.
▪ "course_id" cross-links with the courses table to state which course is
being taught within the class session.

2. Teachers Table

The "teachers" table is responsible for storing detailed information about the teaching
staff. Teachers are an essential part of any educational system, and this table ensures that
their data is managed efficiently.

• Table Structure and Fields:

▪ The "id" field uniquely identifies each teacher in the system.
▪ Personal information such as "username", "first_name", "last_name",
"email", and "password" is stored for authentication and communication
purposes.
▪ Flags like "is_superuser","is_staff", and "is_active" indicate the teacher’s
role, permissions, and active status within the system.
▪ The "date_joined" field logs when the teacher first joined the institution,
and the "last_login" field tracks their most recent activity.

• Foreign Key Relationships:

▪ The "department_id" cross-links to the departments table, identifying to
which department the teacher is associated with.

21
Figure 3.4: SAMS Database Structure

22
 • Relationship to Timetables:

▪ The teachers table is connected to the timetables table through the

"teacher_id" foreign key. This relationship ensures that each teacher can
be assigned to specific classes based on the timetable.

3. TeacherCourses Table

The "teachercourses" table acts as a junction table that facilitates many-to-many

relationships between teachers and courses. In real-world scenarios, a teacher may be
responsible for teaching multiple courses, and a course may be taught by more than one
teacher.

• Table Structure and Fields:

▪ The table contains two primary fields: "teacher_id" and "course_id".

• Foreign Key Relationships:

▪ The "teacher_id" field is a foreign key that references the teachers table.

▪ The "course_id" field is a foreign key that connects to the courses table.

4. Courses Table

The "courses" table stores all the information about the courses offered by the educational
institution.

• Table Structure and Fields:

▪ "course_id" uniquely identifies each course.

23
 ▪ "course_name" provides the name of the course, making it easy for
teachers, students, and administrators to reference the course.

5. Students Table

The "students" table manages detailed records of all enrolled students in the institution. It
is a critical table within the SAMS database, enabling student-related operations like
attendance, reporting, and academic tracking.

• Table Structure and Fields:

▪ Personal information fields include "branch", "roll_no", "full_name",
"email", "date_of_birth" and "mobile_number".

▪ The "gender" field records the student’s gender.

▪ Academic details include "course", "section", "start_year", and

"end_year", which help in tracking a student's academic progress.

▪ The "image" field stores the student’s profile picture, which will be used
for identity verification purposes.

• Foreign Key Relationships:

▪ The "department_id" cross-links to the departments table, identifying to
which department the student belongs to.

6. Departments Table

The "departments" table organizes information about the academic departments within
the institution.

24
 • Table Structure and Fields:
▪ "dept_id" uniquely identifies each department.

▪ "dept_name" provides the full name of the department (e.g., Computer

Science, Mechanical Engineering).

▪ "short" gives a short or abbreviated version of the department name, which

can be used for quick references.

The departments table connects to the timetables table through the "department_id"
foreign key. This relationship ensures that each class is associated with a valid department,
facilitating organized scheduling and departmental reporting.

Key Relationships and Database Integrity

The SAMS database is designed with relational integrity in mind, ensuring that all tables
are interconnected through appropriate foreign key relationships. These connections allow
the system to:

1. Manage teacher assignments efficiently.

2. Schedule classes based on courses, departments, and teachers.

3. Maintain accurate records of student details and enrollment.

4. Generate attendance reports, class schedules, and performance analytics.

25
3.6 System Flow Components

The SAMS system is a system that operates with video feeds captured through cameras
capturing the students in class. The video feeds then undergo processing through the use
of the face_recognition library, which is Python-based and supports real-time face
detection and recognition. This will analyze images of whether there are students in them
and cross reference these real time features of the faces found to those within a pre-existing
database of student faces, thereby enabling it to trace these students from several angles
while the dynamic environment of classroom helps make sure real-time attendance will be
taken appropriately.

This system centrally maintains the attendance data in an attendance database. Teachers
and administrators can check the database for reporting and analyzing attendance. This
whole process eliminates manual attendance taking and increases accuracy without
opportunities to alter attendance data.These are a series of components that take on
individual functions within the process of automating attendance management.

Preprocessing of facial data: Before the system could recognize his face, his facial data
have to be first captured and then processed. To this end, features on every student are
scanned in the registration process, converted into facial embeddings - a digital data that
signifies and represents unique and characteristic qualities of a given student's face. They
can, therefore, be used when the student comes in front of the system and tries to register
his attendance. The fac ial embeddings are stored safely in the database system; they are
not easily accessed or altered and, hence, offer assurance of securing privacy towards
student information. It also cleans the images in the preprocessing so that factors like bad
lighting or image distortion will not affect recognition accuracy. The preprocessing further
retrieves all facial key features like the eyes, nose, and mouth and are used to create a
unique identifier for each student.

26
Real-time Detection and Recognition: The core of the system is its real-time detection and
recognition capability. With the help of face_recognition technology, the system scans the
live feed of the classroom captured by cameras in real-time. It detects the students' faces
entering the classroom and compares them with the facial embeddings saved in the
database. After finding a match, it immediately updates the attendance record by marking
the student present.

This is achieved quite rapidly and non-intrusive in terms of class time such that this process
is inconvenience-free for students and teachers alike. In the system, it even provides an
instance to detect multiple students without restriction even in a completely occupied class
so as to have an accurate count for every student attending.

Management of the attendance database: All data pertaining to the attendance of students
in this database. It is a secure and well-structured database for facial embeddings and
attendance records. The attendance of every student is linked with the facial data, and each
time they are identified, it is logged in the database with timestamps. Teachers,
administrators, or even students can access the database to generate reports, track patterns,
or address attendance-related queries. This will make it easy for the administrators to track
and analyze attendance trends over time.

3.7 Proposed Work

The proposed work is an elaboration of the step-by-step procedures carried out by the
SAMS in its effective automation process for attendance tracking by facial recognition.

1. Enrolling Phase: The enrollment is the starting phase of the operation. It takes facial
data for the students and generates a template to be used in subsequent recognitions.
Facial embedding for all the students with the help of face_recognition library is
obtained with a camera. They are unique for each student, then they are stored
securely within the database. This process will be the following steps:

27
 Face Detection: The system will first detect and locate the face of the student in the
captured image.

Generation of Facial Embedding: The system then converts the face into a
mathematical representation called an embedding that captures the unique features
of the student's face.

Database Storage: The facial embeddings, created are stored in the attendance
database by ensuring that the data concerning the facial images of each and every
student is kept securely safe for further use. At this stage all that is put in for high
quality and secure storage of facial data with proper care and caution for avoiding
breaches of data in order to maintain integrity of the system.

2. Real-Time Recognition This takes place after the enrollment cycle and now the
system gets prepared to recognize the students at once they get inside the classroom.
The use of face_recognition library tries to detect faces using a feed stream from
the camera. When a student gets into the class, then it scans the person's face against
the embeddings stored in the database.

If the system recognizes the student by matching with the face, then it directly
updates the attendance record by marking the student present. This process is done
in real time, and no kind of delay is seen in marking attendance. Therefore, the
system can handle multiple students entering or exiting during the class session,
hence ensuring that attendance is kept updated without interrupting the class at any
stage.

3. End-class verification: In the system, end-class verification would be helpful to

make sure that attendance is actually being captured accurately at any given time
during a class session. The system is responsible for comparing the data collected
on attendance with an expected list of students from the attendees list at the end of
each class session.

28
This also helps in eliminating the students who were only in school for part of a
class, as marks are not awarded entirely. This system checks for entry and exit times
for such pupils and makes appropriate changes. The system has also prevented a
student from leaving early and getting a mark for the whole time of the session.

This phase also supports producing attendance reports indicating not only
attendance but also the period of such attendance. Reports can, therefore, be used
for academic record keeping or distributed to parents and students to enable
transparent record keeping of attendance.

29
 Chapter 4
Implementation

4.1 Introduction

This means that SAMS would use both hardware and software aspects to put together the
totally integrated and working attendance management system. This chapter highlights
how all the constituents of the system are built, tested, and released to a live environment.
For implementation, integration of facial recognition technology, image capture, logging
of attendance, and managing data functions of the system will be utilized.

The system will work in a smooth manner by running the Python libraries, which are
face_recognition and OpenCV, in unison with cameras and servers to ensure real-time
attendance tracking. The purpose of this phase is to align all the components together in a
manner that automates attendance in a way that makes sure the system is efficient and
reliable.

4.2 Implementation Strategy

In the figure 4.1, the algorithm shows the overall Student Attendance Management System
(SAMS) , which uses face recognition technology for automating attendance marking. The
system works in a sequence of steps to ensure accurate attendance recording and reporting.

30
Figure 4.1: Smart Attendance Management System

31
Step 1: Teacher Login

This step provides that only authorized teachers should have access to the system. The
teacher is presented with a prompt to enter their username and password; this is validated
using the `verify_credentials()` function. The system then displays a "Login Successful"
message and continues when the credentials match those in the database. If the entered
credentials are invalid, then it displays an "Invalid Credentials" message and quits to avoid
unauthorized access.

Step 2: Capture Attendance

Once the teacher logs in successfully, the system initializes the camera feed using
`camera_feed`. While the camera feed is active, video frames are continuously captured
using `capture_frame()`. Each frame is processed to detect faces using a face detection
function (`detect_faces`), and the detected faces are appended to the list `captured_faces`.
This step ensures that all visible students' faces are captured for further recognition.

Step 3: Preprocess Captured Images

In this step, the system preprocesses the captured images to enhance the accuracy of the
face recognition process. The function `preprocess_images()` applies image processing
techniques such as converting images to grayscale, noise reduction, and face alignment.
This prepares the detected faces for comparison against the reference database of student
faces.

32
Step 4: Face Recognition and Validation

The preprocessed images of the faces are matched with registered students' face images
present in the database. This is done through a comparison function,
`compare_with_reference()`, which goes through each face image within the list
`processed_faces`. Whenever a face is found, it matches with the face present in the
database, namely `reference_faces`. After matching is successful, a roll number is fetched
accordingly. This system further authenticates the class the recognized student attends.

Step 5: Validate Student

At this point, the system checks if the detected student is valid for the current class by
calling the `is_student_valid()` method. In case of validation success, the attendance status
of the student is updated to "Present" in the `attendance` table. In case the student is not
valid for the class, the system prints a message to show that, and the attendance is not
updated.

Step 6: Mark Attendance

After processing all captured faces, the system iterates through the attendance records using
a loop and updates the attendance table with roll number, current date, attendance status,
and timestamp. This way, the attendance data is kept systematic and accurate.

33
Step 7: Generate Report

Once the attendance is recorded, the system generates a report of attendance by exporting
the `attendance` table into a CSV or Excel file. The report can then be used by the teachers
for documentation or further analysis.

Step 8: Logout

This is the final step where the teacher is logged out, and a confirmation message
"Attendance Completed" is displayed to ensure that the operation has been completed. The
session is closed properly to avoid unauthorized access.

The following will be the core processes constituting the SAMS . These processes are
aimed for the seamless integration of the face recognition technology with attendance
management. These processes work well in a structured way. Here, the first step has been
loading student data with the generation of face encoding, validation of student's
information, and real time processing of attendance. Then, each step would act as a vital
role within the system to ensure efficiency, accuracy, and dependability. These processes
would actually allow the attendance marking automation in a systematic data-handling and
error-free environment easily accessed for reporting through highly developed image
processing techniques.

In figure 4.2, algorithm focuses on loading student images and their corresponding class
details into the system. It initializes the process by accessing the directory media/student
where student image files are stored. It then reads these images and extracts filenames
(without extensions) to populate two lists: images (for storing image data) and classNames
(to store student names or roll numbers).

34
 Figure 4.2: Load Students

The key steps include:

1. Reading files from the specified directory.
2. Loading image data into the images list.
3. Extracting and storing the names or roll numbers from filenames into the
classNames list.
4. Computing known face encodings using the findEncodings() function, which
processes images to generate encodings.
5. Returning encodeListKnown (a list of face encodings) and classNames.

This algorithm ensures the system has preloaded data of all students, which is essential for
face recognition during attendance marking.

35
 Figure 4.3: Encodings

The findEncodings() function is used to process student images and generate face
encodings as shown in figure 4.3. Face encodings are numerical representations of a
person’s face, enabling accurate comparison during recognition.

Steps:
1. Convert each image to RGB format, as most face recognition libraries (e.g.,
OpenCV, Dlib) require RGB images.
2. Compute the face encoding for each image using a face recognition library.
3. Append the encodings to an encodeList.
4. Return the list of face encodings.
5. This algorithm acts as a utility function, ensuring all student faces are encoded for
quick and reliable matching during attendance.

36
 Figure 4.4: Valid Student

The purpose of this algorithm in figure 4.4 is to validate whether a student belongs to a
specific class and section. Given a roll number (roll) and class details (classSec), the
function checks for the student’s existence and verifies their details.

Steps include:

1. Retrieve the student record using the roll number.

2. Validate the classSec format by splitting it into class year, course, and section.
3. Determine the student’s current academic year using the start year and current date.
4. Compare the extracted details with the student’s actual class information.

37
 5. If all details match, return True. Otherwise, return False with an appropriate error
or mismatch message.

This function ensures that only valid students belonging to the specified class and section
are considered for attendance.

Figure 4.5: Attendance

In figure 4.5, the algorithm implements the real-time attendance process using face
recognition and webcam capture. It matches detected faces with preloaded encodings and
marks attendance for students.

38
The steps include:
1. Start webcam capture and continuously process video frames.
2. Resize and convert each frame to RGB format for face recognition.
3. Detect faces in the frame and compute their encodings.
4. Compare the face encodings in the current frame with known encodings
(encodeListKnown) to find the closest match.
5. If a match is found:
• Retrieve the roll number from the classNames list.
• Check if the student belongs to the specified class section.
• Call the markAttendance() function to record the attendance.
6. Display the webcam frame with a bounding box and roll number.
7. Stop the process when the ESC key is pressed or attendance is marked.

This algorithm forms the core of the system by integrating real-time face detection,
recognition, and attendance marking.

Figure 4.6: Mark Automatic Attendance

39
The function in figure 4.6 handles the actual recording of attendance into a CSV file for a
given subject and class section.

Steps include:
1. Validate the subject and class section details.
2. Generate the file path for attendance storage, creating directories as needed
(attendance/date/class).
3. Check for an existing CSV file for the subject. If not present, create a new file
with appropriate headers.
4. Verify if the attendance for the given ID is already marked. If not, append the
student ID and current timestamp to the file.
5. Print a confirmation message to ensure attendance is successfully marked.

This algorithm ensures that attendance data is stored systematically and prevents
duplicate entries.

4.3 Hardware/Tools/Software Requirements

The were various tools, hardware, and software components required to build and deploy
the Smart Attendance Management System (SAMS). The successful and correct
integration of these tools and solutions is the key to the system's functionality and good
efficiency.

• Software Requirements:

1. Python:

It is primarily used to develop the system. Python is chosen for its various
features such as its flexibility, ease of use, and support for extensive and
useful libraries.

40
 Libraries like face_recognition, OpenCV, and SQLite are the most
important for facial detection, video processing, and data management.

2. face_recognition library:

It is a Python library used for facial recognition. It provides easy-to-use

functions for detecting and identifying faces within images and video
frames along with customizability.

It helps in creating facial embeddings and matching them to stored facial

data for real-time identification and reporting.

It has dependencies on CMake and dlib, which are required for compiling
and running the underlying face detection and recognition algorithms
efficiently.

3. OpenCV:

A powerful library for computer vision tasks such as image capture, face
detection, and manipulation of video streams.

OpenCV is used for handling video feeds from the classroom cameras and
processing image frames to extract facial features.

4. SQLite:

A lightweight, file-based database management system used to store

attendance records and facial embeddings securely.

It is chosen for its simplicity, speed, and ability to handle small to medium-
scale data storage, making it ideal for an educational environment.

41
 • Hardware Requirements:

1. Cameras:

High resolution IP cameras are required to capture video footage of students

in the classroom. These cameras are required in order to provide high
resolution images/feed to accurately detect facial features for students.

Placement of cameras should be in a way that it ensures every student’s face

is captured clearly as the class go on along with proper height and angle
alignment.

2. Graphics processing unit:

The computational load for running face recognition algorithms and

managing real-time data may require local servers with adequate processing
power, especially for larger classrooms.

These tools and hardware components are essential to ensure the reliability, scalability, and
efficiency of the system.

4.4 Expected Outcome

The expected outcome of using the Smart Attendance Management System is to achieve a
fully automated, efficient, and accurate attendance system using facial recognition
technology. The following are the primary outcomes that the system aims to deliver.

42
Real time marking of attendance; the attendance is marked in a real time manner as it tracks
how students enter class. Because SAMS scans and recognises faces from students that
enter, updating attendance occurs immediately without awaiting any given time. Taking
attendance automatically obviates manually doing it with the potential errors of having
stale records.

Automated Reporting and Less Human Intervention: SAMS will automatically generate
detailed attendance reports at the end of each class session. There will be no need for
teachers and administrators to waste time manually compiling attendance data because the
system will do all the reporting work. Automation reduces human intervention, meaning
less administrative work, fewer errors, and more time for teaching and student engagement.

43
 Chapter 5
Result & Discussion

5.1 Result

Figure 5.1: Landing Page

Above figure 5.1 depicts the how the website landing page looks like. Faculty can login
from here. Also Above Faculty login is Student registration link which redirects them to
the registration form.

44
Figure 5.2: Registration Form 1

Figure 5.3: Registration Form 2

45
Student Registration form is two-part form as shown in figure 5.2 and figure 5.3. In first
form students would nee d to fill the details and in second part image is captured for face
recognition purpose.

Figure 5.4: Dashboard

After successful login, teachers are redirected to their dashboards where there day's
schedule is given as shown in fig 5.4. In order to take attendance, they will have to choose
the class from the drop and down and click on the start button. When stop button is clicked,
generate report button is appears and on clicking the button students' attendance report is
given for that class and date as shown in figure 5.5. You can download in pdf as shown in
figure 5.6.

46
Figure 5.5: Report Generate View

Figure 5.6: Report PDF Form

47
5.2 Discussion

5.2.1 Initial Tests & Accuracy

The initial tests conducted in the SAMS have shown results with excellent accuracy in
face recognition under strictly controlled conditions. These have been executed in a
fully controlled environment where variables such as lighting, camera angle, and
movement of students will be maintained constant to assess a baseline level of
performance within the system.

Accuracy of Face Recognition: The system had a decent accuracy in recognizing and
matching the faces even if there was an expression on the face or a slight tilt to the
head. Such accuracy is reached using the machine learning algorithm of advanced
degrees and also utilizing the face_recognition library, which utilizes deep learning-
based facial embeddings to guarantee dependable identification.

The preprocessing step is one of the key factors that contribute to the accuracy of the
system, optimizing facial features to ensure that facial embeddings are extracted with
high precision. The system also employs thresholding techniques to ensure that only
clear, legitimate matches are accepted and that false positives or negatives are avoided.

Controlled Conditions: Under controlled conditions, where the classroom environment

is consistent and students' faces are properly captured, the system's performance
remains robust and reliable.

Factors such as camera resolution, facial alignment, and distance from the camera all
played a role in achieving this high accuracy during the tests. These controlled tests
confirmed the effectiveness of the pre-enrollment process and real-time matching in
ideal conditions.

48
 5.2.2 System Scalability

One of the key attributes of SAMS is its ability to scale effectively to accommodate
different classroom sizes and handle varied lighting conditions.

Classroom Size: SAMS has been designed to handle classrooms of varying sizes. The
real-time face detection and recognition algorithms can process multiple students
simultaneously, ensuring that the system works efficiently even in larger classrooms
where many students are present at the same time.

Lighting Conditions: The system's performance was also tested under variable lighting
conditions, ranging from well-lit classrooms to dimly lit ones. Such a system is needed,
as many classrooms are susceptible to fluctuations in light resulting from windows,
artificial lighting, or other environmental factors.

SAMS still performed very well in the low light where it took advantage of applying
open CV's image enhancement algorithm hence helping improve quality in a situation
poorly illuminated. Although the system produced little decreased accuracy in extreme-
low light condition, thus insinuating the system is sensible to too little lighting in some
points.

5.3 Integration Challenges & Mitigation Strategies

While the implementation of SAMS has been largely successful, several integration
challenges were encountered during the development and deployment phases. The
following challenges were identified, along with the strategies employed to address them:

1. Challenge: Hardware Compatibility and Camera Placement: A major challenge was

ensuring that the cameras used for face capture were compatible with the software
and placed effectively in the classroom. In classrooms with high ceilings or large

49
 distances between students and cameras, it was difficult to ensure that faces were
captured clearly from all angles.

Mitigation Strategy: To achieve this, several cameras were mounted in the

classroom in a way that would cover a wide angle of the classroom to ensure that
all students are captured. Additionally, the system was designed to handle the input
from several cameras at once, improving coverage and reducing the chances of
missing some students.

Camera calibration and optimal positioning guidelines were also developed to

ensure that the system operates efficiently regardless of classroom size.

2. Challenge: Real-time Processing Overhead: Handling real-time video processing

for face detection and recognition can be computationally demanding, especially in
large classrooms or when using high-resolution cameras. This can lead to potential
delays or lags in recognizing faces, especially with multiple students entering the
room at once.

Mitigation Strategy: To overcome this problem, the system was optimized by using
multi-threading and parallel processing to process multiple frames simultaneously.

3. Challenge: Handling Multiple Students Simultaneously: In classrooms with many

students, it was challenging to ensure that the system accurately detected and
recognized all students without confusion, especially when students were moving
or entering the classroom at different times.

Mitigation Strategy: To address this, the system was enhanced with dynamic face
tracking, which continuously monitors the classroom for new students and rechecks
faces already detected. This allows for accurate attendance marking even in
situations where multiple students enter the room at once.

50
 Additionally, time-based algorithms were implemented to ensure that students
entering after the initial recognition process could still be logged correctly without
disrupting the flow of data.

4. Challenge: Privacy Concerns and Data Security: Given the sensitive nature of facial
data, privacy concerns were raised regarding the storage and handling of student
facial images and attendance records.

Mitigation Strategy:

The system addresses these concerns by making sure that all facial data is encrypted
and stored with privacy regulations, such as GDPR or local educational data
protection laws.

This data is also accessible to authorized personnel only, with the additional feature
of putting all attendance logs in databases with audit trails for security and
accountability.

51
 Chapter 6
Conclusion & Future Scope

6.1 Conclusion

SAMS eliminates all inefficiencies and the challenges associated with traditional
attendance taking methods, such as high time consumption and errors due to manual
marking. Our system uses facial recognition technology for automatic attendance, bringing
along several key advantages over the conventional methods:

1. Accuracy and Reliability: SAMS has proved to be highly accurate in face

recognition, which minimizes the possibility of proxy attendance and human error.
Since the system depends on facial features unique to a person, it ensures that
attendance records are tamper-proof and reliable.

2. Scalability: The system is scalable to accommodate any size of classrooms and it

adapts to different lighting conditions. It has demonstrated the ability to handle real-
time marking of attendance in a class with up to 50 students, thus proving it can be
deployed both in small classrooms and large lecture halls without significant
performance degradation.

3. Automation and Time Saving: The automation process has significantly reduced
the load of administrative work by freeing teachers or administrators from having
to spend time on attendance-taking. The real-time attendance logging and

52
 automated report generating ensures that teachers have adequate time to teach and
support students, thereby making class operations more efficient.

4. Security and Tamper-Proof: It is one of the basic advantages of SAMS that it

provides tamper-proof nature, unlike traditional systems where students cheated by
marking someone else's attendance. The use of facial recognition ensures that
attendance is recorded when the face of the student gets detected. This improves
accountability and prevents fraudulent marking of attendance.

5. Smooth Integration: The system can smoothly integrate with databases commonly
in use such as SQLite, and uses the face_recognition library combined with
OpenCV to perform image processing, which allows for easy deployment and
maintenance.

Overall, the Smart Attendance Management System not only provides a cost-effective and
efficient attendance-tracking solution but also offers an added level of security and
automation that significantly improves the experience of education for both teachers and
students.

6.2 Future Scope

Although SAMS has laid a strong foundation for an automated attendance tracking system,
there are several areas where it needs further improvement and expansion to make the
system even more robust and adaptable to emerging needs in the educational domain. Some
of the potential future improvements and expansions of SAMS are as follows:

53
Real-Time Student Counting with the Integration of Advanced Machine Learning
Models

Current system uses basic face recognition models from libraries such as face_recognition
that work great under controlled conditions. In the case of real-time student counting in
dynamic environments of classrooms, there is a lot of scope to improve accuracy and
robustness.

Incorporation of the latest deep learning models such as convolutional neural networks
with specific architectures like EfficientNet, DenseNet, or improved variants of ResNet
would deal with issues like obstacles, angle variation, and varying lighting. These models
allow for increased accuracy in student detection and counting simultaneously.

Such improvements will increase the accuracy of counting students in real time and reduce
errors, thereby making the system more reliable in attendance management.

Remote Attendance Tracking for Online Classes

The current system is designed primarily for physical classroom settings, but there is
growing demand for remote attendance tracking in online education or hybrid learning
environments.

Future versions of SAMS could be extended to support remote face recognition through
webcams during virtual classes. This would allow institutions to ensure attendance
accuracy in online courses by detecting students’ faces during live video sessions.

Additionally, integrating the system with existing online learning platforms (e.g., Zoom,
Microsoft Teams, Google Meet) could enable seamless attendance logging and tracking
for remote learners.

54
Integration with Institutional Management Systems

For broader utility, SAMS could be integrated with existing institutional management
systems such as Learning Management Systems (LMS), Student Information Systems
(SIS), and academic portals.

It will then allow attendance data automatically syncing to grade books and other student
profiles as well as other academic records, such that both attendance reports as well as
student performance metrics in real-time are provided from a central dashboard to benefit
both the teacher and administrator.

Such integration with SMS or email systems may also be allowed for automatic
notifications of students or parents about attendance problems such as frequent absences,
thereby improving communication between educators, students, and families.

Better User Interface and User Experience

The system is already functional but can be made more intuitive and even easier to use for
both teachers, students, and administrators by enhancing the user interface and overall UX.

For the future, applications could be developed both for teachers and students where in
real-time information about attendance status may be received, reminders can be sent
regarding the schedule of classes, or a simple check of attendance reports by a button. Web-
based or cloud-based dashboards may be designed so that administrators may access the
system from any location with improved accessibility and management of attendance data.

Advanced Security Features

Even though SAMS has a secure solution in its system, the future would probably add other
security levels to ensure the privacy and integrity of the students’ data.

55
Advanced encryption techniques, such as end-to-end encryption, could be employed to
ensure that facial data and attendance logs are stored securely.

It could also implement multi-factor authentication both for users, such as teachers and
administrators, and for students to ensure that only the authorized individual can access
sensitive data.

Integration with Other Biometrics

Although facial recognition provides a secure solution, combining other biometric

modalities, such as fingerprint recognition, can provide an additional layer of
authentication and redundancy.

This would be especially useful in environments where facial recognition might fail due to
low lighting, camera malfunctions, or other constraints. Combining multiple biometric
systems could ensure robust attendance management in a variety of conditions.

Customizable Features:

Institutions may have different needs based on their size, student population, and specific
requirements. Future versions of SAMS could include more customizable features, such
as: Custom attendance settings (e.g., grace periods for late arrivals), Multi-location support,
where the system can handle attendance tracking across different campuses or classrooms.

56
References

1. Raj, A. A., Shoheb, M., Arvind, K., & Chethan, K. S. (2020, June). Face recognition
based smart attendance system. In 2020 International conference on intelligent
engineering and management (ICIEM) (pp. 354-357). IEEE.
2. Dassanayake, S., & Wanniarachchi, A. (2021, November 15). Challenges of manual
attendance system towards student motivation. ResearchGate.
https://www.researchgate.net/publication/356202082
3. Zhao, M., Zhao, G., & Qu, M. (2022). College Smart Classroom Attendance
Management System Based on Internet of Things. Computational intelligence and
neuroscience, 2022, 4953721. https://doi.org/10.1155/2022/4953721
4. Pereira, H., Carreira, R., Pinto, P., & Lopes, S. I. (2020, June). Hacking the RFID-based
Authentication System of a University Campus on a Budget. In 2020 15th Iberian
Conference on Information Systems and Technologies (CISTI) (pp. 1-5). IEEE.
5. Duwal, R., & Tamang, S. (2025, January 3). Attendance system with facial recognition
using AI. ResearchGate. https://www.researchgate.net/publication/358722010
6. Perwej, Y., Trivedi, A., Tripathi, C., Srivastava, A., & Kulshrestha, N. (2022). Face
recognition based automated attendance management system. International Journal
of Scientific Research in Science and Technology, 9, 261–268. ResearchGate.
https://doi.org/10.32628/IJSRST229147
7. Yao, Y. (2024, January 22). Research on facial recognition system based on deep
learning. Applied and Computational Engineering, 34, 232–236.ResearchGate.
https://doi.org/10.54254/2755-2721/34/20230332

57