VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Belagavi - 590018

A Project Report on

REAL TIME AI POWERED FACE RECOGNITION

SYSTEM WITH DATABASE INTEGRATION
Submitted by
AHANIJA NARAYANA CH 4SN21AD003
PRANAVA SHARMA K 4SN21AD036
PRATHVIRAJ 4SN21AD038
WINSTON FERNANDES 4SN21AD053

In partial fulfillment of the requirements for the degree of

BACHELOR OF ENGINEERING
IN

ARTIFICIAL INTELLIGENCE AND DATA SCIENCE

Under the Guidance of
Prof. NAGARAJA HEBBAR N
Associate professor

Department of Artificial Intelligence and Data Science

SRINIVAS INSTITUTE OF TECHNOLOGY
(ACCREDITED by NAAC, “A” GRADE)
MANGALURU – 574143, KARNATAKA
2024-2025
 SRINIVAS INSTITUTE OF TECHNOLOGY
(ACCREDITED by NAAC, “A” GRADE)
MANGALURU – 574143, KARNATAKA

DEPARTMENT OF ARTIFICIAL INTELLIGENCE AND DATA SCIENCE

CERTIFICATE
Certified that the project work entitled “REAL TIME AI POWERED FACE
RECOGNITION SYSTEM WITH DATABASE INTEGRATION” is a bonafide
work carried out by

AHANIJA NARAYANA CH 4SN21AD003

PRANAVA SHARMA K 4SN21AD036
PRATHVIRAJ 4SN21AD038
WINSTON FERNANDES 4SN21AD053

In partial fulfillment for the award of BACHELOR OF ENGINEERING in

ARTIFICIAL INTELLIGENCE and DATA SCIENECE of the VISVESVARAYA
TECHNOLOGICAL UNIVERSITY, BELAGAVI during the year 2024-2025. It is
certified that all corrections/suggestions for internal assessment have been
incorporated in the report deposited in the department library. The project report
has been approved as it satisfies the academic requirements in respect of Project
work prescribed for the Bachelor of Engineering Degree.

Prof. Nagaraja Hebbar N Prof. Nagaraja Hebbar N Dr. Shrinivasa Mayya D

Project Guide Head of the Department Principal

Name of the Examiners Signature with Date

1.

2.
 ABSTRACT
This project is intended to develop a real-time face recognition system through the
integration of AI with database management for enhanced functionality and performance.
This system detects faces from live video streams using advanced computer vision
algorithms, which thoroughly analyse the faces, drawing out distinctive facial features to
enable deep learning techniques to match faces against an already pre-existing database.
The seamless integration with real-time databases will ensure speedy and accurate
identification, allowing immediate responses based on the recognized individuals.
This system offers a significant advancement in AI-powered computer vision, with
extensive applications in security, access control, and personalized user experiences.
Combined with the latest in AI, this solid database integration provides a system that will
improve not only the field of security but also ensure seamless, customized user
experiences within all manners of environments. There needs to be further development of
these systems in varieties for the continued growth of AI technologies that solve realistic
problems innovatively.
 ACKNOWLEDGEMENT
The success and final outcome of this project required a lot of guidance and assistance from
many people and we are extremely fortunate to have got their support all along the
completion of our project.
We express our heartfelt gratitude to our Parents for their unwavering moral
support, encouragement, and belief in us and motivating us in completing this project.
We extend our profound gratitude and deep regards to our Project Guide, Prof.
Nagaraja Hebbar N, Associate Professor and Head of the Department of Artificial
Intelligence and Data Science, for his invaluable support, constructive feedback, and
guidance throughout the various stages of this project. His enthusiasm, insightful ideas, and
critical reviews have significantly enhanced the quality of our work.
Our sincere thanks go to the Project Coordinator, Prof. Ganapathi Sharma K, Associate
Professor, Department of Artificial Intelligence and Data Science, for his exemplary
guidance, monitoring, and constant encouragement throughout the course of this project.
His contribution will always be remembered.
We also extend our heartfelt gratitude to our Principal, Dr. Shrinivasa Mayya D,
for his kind cooperation and encouragement, which greatly helped us in the completion of
this project.
We express our appreciation to the Management for their direct and indirect
support, ensuring the successful completion of this project.
We are thankful to all the Teaching and Non-teaching staff of the Department of
Artificial Intelligence and Data Science for their constant encouragement, support, and
guidance during this project.
We would like to thank our Friends, with whom we shared our day-to-day
experiences, and who offered invaluable suggestions that improved the quality of our work.
Their companionship and ideas were instrumental in achieving this milestone.

AHANIJA NARAYANA CH
PRANAVA SHARMA K
PRATHVIRAJ
WINSTON FERNANDES
 TABLE OF CONTENTS
CHAPTER TITLE PAGE
No. No.
1 INTRODUCTION 1

1.1 Problem Statement 1

1.2 Proposed System 2

1.3 Project Scope 2

2 LITERATURE SURVEY 4

3 SYSTEM REQUIREMENT SPECIFICATION 7

3.1 Functional Overview 7

3.2 Operating Environment 7

3.2.1 Software Requirements 7

3.2.2 Hardware requirements 8

3.3 Functional Requirements 8

3.4 Non - Functional Requirements 8

3.5 Performance Requirements 9

4 SYSTEM DESIGN 10

4.1 Use Case Diagram for Real Time Face

10
Recognition System with Database Integration

4.2 Sequence Diagram for User and Admin 11

4.3 Data Flow Diagram 12

4.4 Activity Diagram 14

4.5 E-R Diagram 15

4.6 Schema Diagram 16

4.6.1 System Workflow 17

5 SYSTEM IMPLEMENTATION 18

5.1 Algorithm for Real-time AI-powered Face

19
Recognition System with Database Integration

6 SYSTEM TESTING 20

6.1 Testing Methodologies 20

6.2 Testing Criteria 21

7 SCREENSHOTS 22

8 RESULT ANALYSIS 28

9 USER MANUAL 29

CONCLUSION AND FUTURE SCOPE 31

REFERENCES
 LIST OF FIGURES
FIGURE TITLE PAGE
No. No.
4.1 Use Case Diagram for Admin and User 10

4.2 Sequence Diagram for Face Recognition 12

Attendance Management

4.3 Data Flow Diagram for Face Recognition 13

Attendance Management

4.4 Activity Diagram of Facial Recognition 14

4.5 E-R Diagram for Face Recognition Attendance Management 16

4.6 Schema Diagram for Face Recognition 17

Attendance Management

7.1 Register and Login Page 22

7.2 Register Details Page 23

7.3 Home Page 23

7.4 Student Details Page 24

7.5 Student Photo Training Page 24

7.6 Student Face and Spoof Detector Page 25

7.7 Student Attendance Page 26

7.8 Developer Page 26

7.9 Web Camera Implemented to Classroom 27

8.1 Accuracy and Validation of Real-Time AI-Powered 28

Attendance System

9.1 Admin Panel 29

9.2 Attendance Management Panel 29

9.3 Student Panel 29

9.4 Data Train Panel 30

9.5 Face Detector Panel 30

9.6 Attendance Panel 30

9.7 Exit Panel 30

 LIST OF TABLES
FIGURE PAGE
TITLE
No. No.
Test cases for Real-Time AI-Powered Face Recognition
6.1 22
System with Database Integration.
 CHAPTER 1

INTRODUCTION
Real-Time Facial Attendance System Introduction

CHAPTER 1

INTRODUCTION

In the rapidly evolving landscape of educational technology, real-time AI-powered face

recognition systems are emerging as transformative tools for automating attendance
management. By integrating sophisticated computer vision algorithms and artificial
intelligence, these systems streamline the attendance tracking process, providing
educational institutions with a robust solution that enhances efficiency, security, and
accuracy. This project aims to develop a comprehensive face recognition system that not
only captures and verifies student identities in real time but also incorporates advanced
features such as anti-spoofing measures, multi-face attendance marking, and effective data
logging. At the core of our system is the ability to utilize cameras strategically placed in
classrooms and entry points to capture facial images of students. These images are cross-
referenced with a pre-existing database, which includes student photos, ensuring that
attendance records are updated instantly and accurately. This seamless integration
eliminates the traditional challenges associated with roll calls and manual sign-ins, which
are often time-consuming and prone to human error.
The advantages of implementing a real-time face recognition system extend well
beyond the mere convenience of automated attendance. One of the most critical benefits is
the enhancement of campus security. By ensuring that only authorized individuals are
present, educational institutions can better safeguard their environments. This is
particularly vital in today’s context, where safety concerns are paramount.
Additionally, the system generates valuable insights into student attendance
patterns. By analyzing attendance data, educators and administrators can identify trends
related to absenteeism and engagement. This early detection allows for timely
interventions, helping to address potential academic issues before they escalate. The
digitalization of attendance records not only reduces the reliance on paper-based systems
but also minimizes the potential for errors, ultimately fostering a more organized and
efficient administrative environment.

1.1 Problem Statement

Traditional methods of student attendance tracking in educational institutions are
increasingly seen as inefficient and error-prone. Roll calls can consume significant
instructional time, and manual sign-ins are vulnerable to inaccuracies and misuse.
Department of AI & DS, SIT, Mangaluru Page 1
Real-Time Facial Attendance System Introduction

Moreover, existing systems often lack robust security measures, making it difficult to
ensure that only registered students are present on campus. There is an urgent need for an
automated, accurate, and secure system that can effectively streamline attendance tracking
processes and enhance overall institutional security.

1.2 Proposed System

Our enhanced solution, through Python's extensive libraries and algorithms, provides one
such robust real-time face recognition system that is pragmatic for student attendance
management. Ensuring the usability and efficiency of the system by including OpenCV's
powerful image processing along with Tkinter's flexible GUI framework, optimized
grayscale conversion with Python Imaging Library makes the system further accurate,
enabling reliable light recognition in various scenarios involving it.
Central to the effectiveness of our system are several advanced algorithms. Local
Binary Patterns Histograms (LBPH) is particularly adept at recognizing faces under varying
lighting conditions and facial expressions, ensuring dependable identification.
Additionally, Convolutional Neural Networks (CNNs), such as Residual Neural Networks
(ResNets), significantly enhance the accuracy of facial recognition through deep learning
techniques, allowing the model to continuously improve its capabilities over time.
The system also incorporates Haar Cascades and Dlib for superior facial detection
and feature extraction, providing an added layer of security and operational efficiency. A
unique feature of our solution is its ability to mark attendance for multiple faces
simultaneously, automatically logging the time and date of each identification. This
capability not only increases the speed of attendance processing but also ensures that no
students are overlooked.
Furthermore, anti-spoofing measures are integrated to prevent fraudulent attempts
at bypassing the system. By distinguishing between genuine facial data and potential
threats, our solution ensures that the integrity of the attendance process is upheld.

1.3 Project Scope

This project is designed to develop a real-time face recognition system in Python,
effectively integrating technologies such as OpenCV, Tkinter, LBPH like ResNet and Dlib
for comprehensive automated student attendance management. The focus will be on:
 Automating Attendance Tracking: Streamlining the process to minimize time
spent on attendance and maximize instructional time.
Department of AI & DS, SIT, Mangaluru Page 2
Real-Time Facial Attendance System Introduction

 Enhancing Security: Implementing robust measures to ensure only authorized

individuals access educational environments.
 Improving Operational Efficiency: Reducing administrative burdens and
potential errors associated with traditional attendance methods.
This AI-powered face recognition system revolutionizes attendance tracking in real time
for educational institutes. It automates the process while improving operational efficiency
and security. Our project stands as a pioneering solution for modern educational needs.

Department of AI & DS, SIT, Mangaluru Page 3

 CHAPTER 2

LITERATURE SURVEY
Real-Time Facial Attendance System Literature Survey

CHAPTER 2

LITERATURE SURVEY

This paper [1], has proposed a CNN and Adaboost fusion model for multiface recognition
in the automated student attendance system. Haar-cascade-based feature extraction makes
the recognition of more than one face under varying conditions possible, with an accuracy
of 100%. The model incorporates IoT technology in providing real-time global attendance
data collection and exhibits its capability to efficiently track attendance.
In another paper [2], an intelligent face recognition system is proposed for good
attendance management. The approach is said to achieve an accuracy rate of 82.1%, with a
very small processing time. Therefore, the method can be implemented in combination with
other ones and also takes the record-keeping process onto the cloud, hence making it easier
to trace the attendance.
The paper [3], also introduces an automated tracking system through facial
recognition technology, in which features extracted using deep learning combined with
high-end algorithms for identification. It will track attendance at an educational and
corporate level with much efficiency and security to enrich big accuracy through real-time
updating of data.
The Paper [4], proposed a smart attendance system that applies the "Face
Recognition" method for automated creation of attendance. It creates a unique face
encoding from the students and automatically marks attendance that can be viewed through
real-time live video feeds. In this way, all these stored records are maintained on Firebase
Real-time Database in an easily accessible way by any faculty member.
The study proposes in the paper [5], a face recognition system using deep learning,
Principal Component Analysis, Support Vector Machines, and K-Nearest Neighbors to
ease attendance in institutes. This face recognition system has achieved 96.8% accuracy,
further smoothing the process of attendance tracking by reducing frauds and enhancing
accuracy.
This paper [6], proposed the use of OpenCV for automating the attendance record,
presenting an automated face name list. The proposed system allows attendance reporting;
the format for saving is Excel. Besides, it has a high level of recognition accuracy across
higher recognition parameters. These are bound to assist conventional methods of
managing classrooms, largely reducing manpower and lowering expenses incurred during
the process.

Department of AI & DS, SIT, Mangaluru Page 4

Real-Time Facial Attendance System Literature Survey

The paper [7], is realized in this study by a facial recognition-based attendance management
system, which utilizes the Fisher Face Recognizer algorithm and a classifier called Haar
cascade to identify the student and record his attendance. It will also generate and send an
email with the attendance report to professors, automating this task, and thereby enhancing
the efficacy and efficiency of current attendance management practices.
This paper [8], presented an AI-based Face Recognition Attendance System that
allows automation of processes for face recognition to maintain student attendance records.
Such a method strengthens the circle of accuracy in the maintenance of attendance records,
enables efficiency in real-time management and processing, and shrinks administration
overhead. This optimally manages the process proper attendance in real-time, with updated
knowledge and spotless records [9].
The study of paper [10], involves the introduction of a face recognition attendance
system that automates attendance marking using face detection in order to increase the
accuracy in organizational settings. Since they replace manual mechanisms, it actually
saves time and paperwork, hence giving better efficiency and effectiveness related to
attendance systems.
The paper [11], proposes a Facial Recognition Attendance Monitoring System,
which uses deep learning for image capturing and face detection through algorithms such
as Haar-cascade and Fisher Face, followed by marking the attendance in Excel sheets. This
method enhances accuracy and efficiency significantly in educational settings.
This paper [12], describes a real-time face recognition attendance system based on
the usage of facial recognition technologies. The proposed system automates attendance
and eliminates the use of proxy attendance in workplaces and institutes for higher learning
by incorporating an already trained SVM classifier using CNN for feature extraction.
Another paper [13], application could be the attendance system of the students
through face recognition. This will automatically update student attendance in real-time
without requiring human intervention to count each person. This has the ability to use
OpenCV, Dlib, and CNN to detect and recognize faces; hence, marking attendance fast
even in a highly crowded environment.
The paper [14], describes a real-time attendance system with 92% accuracy in
perfect lighting. For face detection, OpenCV has been used, while for recognition, KNN is
applied. It captures ten images of every individual so that the system can be robust enough
for different facial expressions.

Department of AI & DS, SIT, Mangaluru Page 5

Real-Time Facial Attendance System Literature Survey

In the paper [15], further presented the advantages of the Facial Recognition-based
Attendance System, embodied with anti-spoofing techniques. These systems enhance
precision and security in attendance tracking of both educational and corporate levels
through advanced algorithms tracing special facial features.
In paper [16], proposes a novel face anti-spoofing approach using 3D convolutions
combined with vision transformers for better liveness detection. Besides, learnable gradient
operators are incorporated and achieve lower classification error rates compared to
traditional 2D CNN methods on several datasets.
In the paper [17] and proposes a lightweight face recognition attendance system
with liveness detection using MobileNetV2 on a Raspberry Pi. This system gives 96%
accuracy for live subjects with processing times less than 0.6 seconds and effectively
reduces face spoofing attacks while ensuring portability.

Department of AI & DS, SIT, Mangaluru Page 6

 CHAPTER 3

SYSTEM
REQUIREMENT
SPECIFICATION
Real-Time Facial Attendance System System Requirement Specification

CHAPTER 3

SYSTEM REQUIREMENT SPECIFICATION

The fundamental purpose of System Requirement Specification is to turn a customer's goals

into a document that may be accepted. The customer clearly states what and all they need
from the produced approach, and the developer clearly understands what competency is
necessary to build the system through System Requirement Specification. It consists of a
number of aspects that attempt to explain the customer's planned functionality in order to
please their varied consumers. The purpose of this document is to act as a guide for
developers and testers who are in charge of the system's expansion.
In the proposed system, different human faces are prepared to form an image
dataset. Features like facial shape, size, symmetry, skin tone, etc., are extracted that will
train the system. Feature extraction plays a major role in image processing, as these features
will describe various characteristics of a face that must be considered by the system before
the final classification or recognition. The backend does internal processing through a pre-
trained model when it gains an input provided by the user, which may predict or identify
the person and integrate with a database for further actions.

3.1 Functional Overview

 The user provides an input image of a human face.
 An algorithm is used to train the model with facial features.
 The result is displayed, identifying the individual and integrating it with the
database.

3.2 Operating Environment

The operating environment requires the system to have minimum software and hardware
requirements.
3.2.1 Software Requirements
 Python: Version 3.12.6 for the development and execution of the face recognition
system.
 OpenCV: Version 4.10.0 Essential for image processing, facial detection, and
feature extraction.

Department of AI & DS, SIT, Mangaluru Page 7

Real-Time Facial Attendance System System Requirement Specification

 Tkinter: Version 8.6, a standard library for developing the user interface in Python.
 Python Imaging Library (PIL): Used for image pre-processing, particularly for
Gray scaling and enhancing image quality.
3.2.2 Hardware Requirements
 Processor: Minimum Intel Core i5 or equivalent AMD processor to handle
computational demands efficiently.
 Storage: A minimum of 16 GB RAM and 512 GB SSD to support smooth operation
and data storage.
 Web Camera: A camera with a minimum resolution of 12 MP to capture high-
quality facial images.
 Computer: It act as a both input and output device having above mentioned
features, receiving data through peripherals like keyboard and mouse and delivering
result via system GUI.

3.3 Functional Requirements

A device's or a component feature's functionality is determined by functional criteria. A
function is defined as a set of inputs, actions, and outputs. System-specific outcomes are
defined by functional needs. The application architecture of a system is driven by functional
requirements. The following are the functional requirements used in the project:
 Process images in various formats, such as jpg, png, jpeg, etc.
 Import and store images in the database.
 Perform segmentation and feature extraction of facial attributes.
 Identify the individual based on facial recognition and link it to the database.

3.4 Non-Functional Requirements

Non-functional requirement is a requirement that specifies standards that can be utilized to
determine how a system operates rather than specific behaviours. Non-functional
requirements are a term used to describe a system's quality attributes. The application's non-
functional requirements are listed below.
 Availability: The device should operate according to the stated requirement as
necessary.
 Reliability: The System has the ability to consistently perform the intended or required
with the prescribed requirements.

Department of AI & DS, SIT, Mangaluru Page 8

Real-Time Facial Attendance System System Requirement Specification

 Maintainability: In accordance with this, maintenance of a functional unit can be

performed.
 Accessibility: The System can be accessed by any appropriate users.
 Operability: The System has the ability to keep itself safe and reliable, according to
predefined operational requirements.
 Usability: The System is ease of use and learnable by the Users.
 Responsiveness: The system has ability to respond to the user very fast as soon as the
input is fed.
 Cost: The system is economically feasible.

3.5 Performance Requirements

This face recognition for attendance in real time, using AI for target educational institutes,
is the ultimate face recognition system. Advanced technology integrated not only automates
the process of tracking attendance but also enhances operational efficiency and security,
hence making the learning environment much safer and organized. While each institution
gradually shifts towards complete digital transformation, our project places itself as one of
the pioneering solutions to meet the demand in modern education.

Department of AI & DS, SIT, Mangaluru Page 9

 CHAPTER 4

SYSTEM DESIGN
Real-Time Facial Attendance System System Design

CHAPTER 4

SYSTEM DESIGN

4.1 Use Case Diagram for Real Time Face Recognition System
with Database Integration
The use case diagram models the interactions between the actors (users) and the system for
real-time face recognition attendance. It highlights the roles of the student and the
administrator in managing and tracking attendance via the system.
 System Overview:
o The system involves capturing a student’s image and verifying their identity
through face recognition, enabling them to either register a new face or take
attendance.
 Actors:
o Student: Initiates the system by clicking “Take image.” If recognized, the
system marks attendance; otherwise, the student may register as a new user.
o Administrator: Monitor’s attendance reports generated by the system for
efficient tracking and record management.
 Use Cases:
o Register New Student: Allows students to input their face into the system for
future recognition.
o Take Attendance: Matches the face with existing data and records the student's
attendance.
o View Attendance: Admins can access and manage attendance reports based on
recognized faces.

Figure 4.1: Use Case Diagram for Admin and User

Department of AI & DS, SIT, Mangaluru Page 10

Real-Time Facial Attendance System System Design

Figure 4.1 showcases a face recognition system designed for student attendance. We see a
student interacting with the system, represented here as a computer. To begin, the student
clicks "Take image," prompting the system to identify their face.
Following successful face detection, the student has two options. They can either
register a new face by clicking "Register new students" and taking another picture. which
will be stored in the system's database for future recognition. Alternatively, if the student
is already registered, they can choose "Take attendance." The system then compares the
detected face to the existing database and marks their attendance accordingly.
An administrator also plays a role in this system. They have access to view
attendance reports generated by the face recognition software, allowing them to monitor
student presence efficiently.

4.2 Sequence Diagram for User and Admin

The sequence diagram provides a detailed view of the steps involved in the face recognition
attendance process, including the sequence in which these actions occur between different
components.
 Actors:
o Student: The main user of the system whose face is captured for attendance.
o System Components: Camera, Pre-processing unit, Database, and Feature
Extraction module.
 Process Flow:
o Face Capture: The camera captures an image of the student's face.
o Pre-processing: The system processes the image to remove noise and enhance
features.
o Feature Extraction: Key facial features are extracted to form a unique
mathematical. representation.
 Face Matching:
o The extracted features are compared with the database.
 Confirmation:
o If the face is recognized, the system records attendance.
o If the face is not recognized, it is flagged as unknown, prompting the student
for further action (e.g., registration).

Department of AI & DS, SIT, Mangaluru Page 11

Real-Time Facial Attendance System System Design

Figure 4.2: Sequence Diagram for Face Recognition Attendance Management

This figure 4.2 visualizes the interactions and the order in which the system components
interact with the student, leading to either successful recognition and attendance marking
or further action for unrecognized leading to either successful recognition and automated
attendance marking by the system, or triggering a manual intervention for unrecognized
students.

4.3 Data Flow Diagram

Data flow diagram (DFD) is a graphical representation of the “flow” of data through an
information, modelling its process aspects. Often, they are preliminary steps used to create an
overview of the system. DFDs can also be used for the visualization of data processing
(structured design). Dataflow diagram are constructed from a limited repertoire of shapes,
connected with arrows. Figure 4.3 shows dataflow for Real Time Face Recognition System
with Database Integration.

Department of AI & DS, SIT, Mangaluru Page 12

Real-Time Facial Attendance System System Design

Figure 4.3: Data Flow Diagram for Face Recognition Attendance Management

1. Start: The process begins when the system is turned on.

2. Input Image: The system prompts the user to input an image. This image is typically
captured by a camera connected to the system.
3. Face Detection: The system’s software then uses facial detection algorithms to
identify if there is a face present in the image.
4. Face Recognition: If a face is detected, the system attempts to recognize the face by
comparing it to the faces stored in its database.
 Unknown Faces: If the system cannot recognize the face, it determines this
as an unknown face. The system may then prompt the user for further
instructions, such as if they want to register the new face.
 Known Faces: If the system successfully recognizes the face, it proceeds to
the next step.
5. Take Attendance/ Save Face (Depending on Recognition):
 Known Faces: For recognized faces (i.e., registered students), the system
marks their attendance. It might also save a record of the student’s
attendance along with a timestamp.
 Unknown Faces: If the face is unrecognized, the system may allow the user
to register their face by saving the image in the database.
6. Exit: After completing the above steps, the process ends. The system may wait for
a new image to be input or it might shut down.

Department of AI & DS, SIT, Mangaluru Page 13

Real-Time Facial Attendance System System Design

4.4 Activity Diagram

An activity diagram visually presents a series of actions or flow of control in a system
similar to a flowchart or a data flow diagram. Activity diagrams are often used in business
process modelling. They can also describe the steps in a use case diagram. Activities
modelled can be sequential and concurrent. In both cases an activity diagram will have a
beginning and an end.

Basic activity diagram notations and symbols:

 Rectangles: Rectangles are used to represent actions or steps in a process. In the

image, the rectangles say things like "Face Detection" and "Feature Extraction"
which are all steps involved in identifying a face using facial recognition.
 Arrows: Arrows show the flow of the process. In the image, arrows connect the
rectangles showing the order the steps are completed. For example, an arrow goes
from "Camera" to "Face Detection" showing that the image captured by the camera
is used in the face detection step.
 Swimlanes: Swimlanes are used to group related activities together. While not
explicitly labeled in the image, the separation between "Training" and "Testing"
sections acts functionally like swimlanes.

Figure 4.4: Activity Diagram of Facial Recognition

1. In this figure 4.4, the system is divided into two sections: Training and Testing. The
Training section refers to how the system learns to recognize faces, while the
Testing section refers to how the system is used to identify faces in an image or
video.

Department of AI & DS, SIT, Mangaluru Page 14

Real-Time Facial Attendance System System Design

2. In the Training section, a dataset of images is used to train the system. This dataset
likely consists of multiple images of each person the system needs to recognize.
Each image is labelled with the identity of the person pictured.
3. During training, the system extracts feature from the training images. Facial features
are mathematical representations of a face’s characteristics, like the distance
between the eyes or the shape of the jawline.
4. A machine learning model, such as a Support Vector Machine (SVM), is used to
classify the extracted features. In essence, the SVM learns to associate specific
features with specific identities.
5. In the Testing section, a camera captures an image or video of a person. The system
then extracts features from this image and uses the SVM to classify those features,
identifying the person in the image or video.

4.5 E-R Diagram

An entity relationship (ER) diagram shows the relationships of entity sets stored in a
database. An entity in this context is a component of data. In other words, ER diagrams
illustrate the logical structure of databases. At first glance an entity relationship diagram
looks very much like a flowchart. It is the specialized symbols, and the meanings of those
symbols, that make it unique.

Common Entity Relationship Diagram symbols:

 Entities, which are represented by rectangles. An entity is an object or concept about
which you want to store information.
 Actions, which are represented by diamond shapes, show how two entities share
information in the database.
 Attributes, which are represented by ovals. A key attribute is the unique,
distinguishing characteristic of the entity. For example, an employee's social
security number might be the employee's key attribute.
 Connecting lines, solid lines that connect attributes to show the relationships of
entities in the diagram.
 Cardinality specifies how many instances of an entity relate to one instance of
another entity. Ordinality is also closely linked to cardinality. While cardinality
specifies the occurrences of a relationship, ordinality describes the relationship as
either mandatory or optional.

Department of AI & DS, SIT, Mangaluru Page 15

Real-Time Facial Attendance System System Design

 Total participation – Each entity is involved in relationship. Total participation is

represented by double lines.
 Partial participation – In this not all entities are involved in the relationship. Partial
participation is represented by single lines.

Figure 4.5 shows the E-R diagram for vehicle document management. There are 3 entities
and they are User, Student and Attendance. The relation between User and Student is M:
N. The relation between User and Attendance is 1:N with partial participation and same
between Student and Attendance.

Figure 4.5: E-R Diagram for Face Recognition Attendance Management

4.6 Schema Diagram

A database schema is the skeleton structure that represents the logical view of the entire
database. It defines how the data is organized and how the relations among them are
associated. It formulates all the constraints that are to be applied on the data. A database
schema defines its entities and the relationship among them. It contains a descriptive detail
of the database, which can be depicted by means of schema diagrams. It’s the database
designers who design the schema to help programmers understand the database and make
it useful.

Department of AI & DS, SIT, Mangaluru Page 16

Real-Time Facial Attendance System System Design

This schema diagram emphasizes the streamlined process of student registration, face
recognition, and attendance marking, while also facilitating administrative oversight
through an efficient database management system.

Figure 4.6: Schema Diagram for Face Recognition Attendance Management

Figure 4.6, illustrates the interaction between the student and the administrator within the
face recognition attendance system. The student is depicted as a stick figure, and the admin
is represented by a computer.

4.6.1 System Workflow

1. Student Interaction: The process begins when a student initiates the system by
clicking “Take Image.” At this point, the system activates its face detection
functionality.
2. Face Detection: Once the system detects a face, the student is given two options:
 Register New Face: If the student chooses to register a new face, they can click
“Register New Students.” A new image will be captured, which is then stored
in the system’s database for future recognition.
 Take Attendance: If the student is already registered, they can opt to mark their
attendance by choosing “Take Attendance.” The system compares the detected
face with those stored in its database. If the face is recognized, the system
records the student's attendance.
3. Admin Access: The administrator has the authority to view and manage the
attendance reports generated by the system, ensuring accurate tracking and
monitoring of student attendance.

Department of AI & DS, SIT, Mangaluru Page 17

 CHAPTER 5

SYSTEM
IMPLEMENTATION
Real-Time Facial Attendance System System Implementation

CHAPTER 5
SYSTEM IMPLEMENTATION
System implementation is the stage where the theoretical design is converted into a working
system, the new system may be totally new, replacing an existing manual, or automate
system or it may be a major modification to an existing system. The project is implemented
using MySQL Workbench and Visual Studio.

5.1 Algorithm for Real-time AI-powered Face Recognition

System with Database Integration
Step 1: Start
 Initialize the camera or video feed.
 Load the pre-trained deep learning model for face recognition (e.g., VGGFace2,
FaceNet).
 Connect to the database containing registered faces and their associated
information.
Step 2: Continuous Loop
 Capture a frame from the camera/video feed.
Step 3: Face Detection
 Use a pre-trained face detection model (e.g., Haar cascade classifier) to identify
faces within the captured frame.
 If a face is detected, proceed to Step 4.
 If no faces are detected, go back to Step 2 and continue capturing frames.
Step 4: Face Preprocessing
 Crop the detected face region from the frame.
 Resize the cropped image to the size expected by the face recognition model.
 Apply any necessary normalization techniques (e.g., color normalization) to the
image.
Step 5: Feature Extraction
 Pass the preprocessed image through the pre-trained face recognition model.
 The model will extract a facial feature vector representing the unique
characteristics of the detected face.

Department of AI & DS, SIT, Mangaluru Page 18

Real-Time Facial Attendance System System Implementation

Step 6: Face Recognition

 Compare the extracted feature vector with the feature vectors stored in the database
for registered faces.
 Use a similarity metric (e.g., cosine similarity) to measure the distance between the
extracted vector and each stored vector.
Step 7: Recognition Result
 Match Found (Known Face):
o Identify the registered user based on the closest match in the database.
o Display the user's information on the screen (optional).
o Depending on the application, trigger further actions (e.g., unlock a door,
grant access, record attendance).
 No Match Found (Unknown Face):
o Inform the user that the face is not recognized.
o Optionally, provide functionalities for new user registration (capture
additional data and store features with user information).
Step 8: Database Integration
 Update the database with any relevant information about the detected face (e.g.,
timestamps, access attempts for unknown faces).
Step 9: Go back to Step 2
 Continue capturing frames, repeating the process to detect and recognize faces in
real-time.
Step 10: End
 This step is typically triggered by the user closing the application or shutting down
the system.
 Release resources like the camera and database connection.

Department of AI & DS, SIT, Mangaluru Page 19

 CHAPTER 6

SYSTEM TESTING
Real-Time Facial Attendance System System Testing

CHAPTER 6

SYSTEM TESTING

Software testing is a common approach for evaluating the accuracy, completeness, security,
and consistency of computer applications. This covers the method by which the system is
run with the goal of detecting errors. Quality is subjective and valuable to different
stakeholders. Testing cannot fully demonstrate the correctness of an arbitrary computer
program but is essential for identifying program errors. Test cases are crucial for exposing
program issues, ensuring that code functions as intended.
Testing is performed to locate flaws, ensuring all previous steps' errors are resolved.
The system runs with a series of test cases to analyse software performance and confirm
proper execution. Identifying faults within a system is the first step toward improving its
functionality. Success depends on research systems revealing these faults effectively.

6.1 Testing Methodologies

The following testing methodologies were employed:

 Unit Testing: Individual components or modules of the face recognition

system, such as image pre-processing and feature extraction, were tested to
verify their correctness.
 Integration Testing: Modules like face detection, recognition, and database
interaction were combined into subsystems and tested to ensure seamless
integration.
 System Testing: The complete system was tested against the requirements to
ensure it met functional and non-functional specifications.
 Acceptance Testing: The system was tested with real-world data to
demonstrate its functionality to stakeholders and ensure it met their needs.

Department of AI & DS, SIT, Mangaluru Page 20

Real-Time Facial Attendance System System Testing

6.2 Testing Criteria

Table 6.1: Test Cases for Real-Time AI-Powered Face Recognition System with
Database Integration.
Sl. Test Procedure Pre-Condition Expected Result Passed/
No. Failed

1 Detect a face in Ensure the camera is The system detects and

a live video functional and properly outlines the face Passed
stream positioned
2 Recognize a Provide a registered The system identifies only
registered face user’s face in front of the registered students. Passed
the camera
3 Detect an Provide an The system shows
unregistered unregistered face in unknown person Passed
face front of the camera
4 Mark Provide a registered Attendance is marked and
attendance for a face during the logged in the database Passed
recognized face attendance-taking
session
5 Prevent spoof Present a photograph Present a photograph or
attempts or video of a registered video of a registered user Passed
user
6 Export Request for an A CSV file with
attendance attendance report from attendance data is Passed
records the admin panel generated

The designed table 6.1, outlines the Test Cases for the Real-Time AI-Powered Face
Recognition System with Database Integration, detailing the Test Procedure, Pre-
Condition, Expected Result, and whether the test Passed/Failed.

Department of AI & DS, SIT, Mangaluru Page 21

 CHAPTER 7

SCREENSHOTS
Real-Time Facial Attendance System Screenshots

CHAPTER 7

SCREENSHOTS

7.1 Register and Login Page

The figure 7.1 describes an important entry point to the real-time AI-powered attendance
system that is designed to deliver secure and efficient access to its major users. On the
register and login page, it streamlines the experience of a user by making available features
for account registration, login, and even password recovery, ensuring easy access and
protection of data. Reflecting the innovation that this system brings forward, it bridges
advanced AI technology with intuitive design for optimizing real-time management of
attendance records for organizations.

Figure 7.1: Register and Login Page

7.2 Register Details Page

The figure 7.2 represents Register details page is the portal through which the new user will
enrol themselves in the real-time AI-Powered attendance system. It presents a user-friendly
interface that reduces the complexity in registration by filling up vital details like first name,
last name, contact number, and the email address. Security Question and Answer are also
present for account protection that ensure privacy about the data and lock down the access.
The page also provides a clear checkbox for accepting the system's terms and conditions,
which is transparent and understood by the user. By combining user-friendly design with
robust security measures.

Department of AI & DS, SIT, Mangaluru Page 22

Real-Time Facial Attendance System Screenshots

Figure 7.2: Register Details Page

7.3 Home Page

The figure 7.3 describes the real-time AI-Powered attendance system's Home Page, which
greets the users with a clean, intuitive interface. It will be the central hub of all attendance-
related activities. Such key features as real-time attendance tracking, employee profiles,
and customizable reports are accessible through appealing icons or menu options. The
Home Page is also a good source of current attendance statistics, up-to-date events, and
even system announcements, giving users the easy view of important information at a
glance. By keeping it user-centric with its accessibility in mind, the Home Page has set the
stage for a more efficient and enjoyable experience within the attendance system.

Figure 7.3: Home Page

7.4 Student Details Page

The figure 7.4 depicts Student Details Page within the real-time AI-Powered attendance
system provides a comprehensive view of individual student information. The page allows
for easy navigation and filtering, enabling users to search for specific students based on

Department of AI & DS, SIT, Mangaluru Page 23

Real-Time Facial Attendance System Screenshots

criteria such as Student ID, Name, Department, Course, Year, and Semester. The "Class
Student Information" section displays detailed personal information including Student ID,
Name, Class Division, Roll No., Gender, DOB, Email, Mobile Number, Address, and
Teacher Name. The page also includes features for photo management, allowing users to
take, update, or replace student photos. With its organized layout and easy-to-navigate
interface.

Figure 7.4: Student Details Page

7.5 Student Photo Training Page

The figure 7.5 depicts Student Photo Training Page in the live AI-Powered attendance
system that is meant to be able to train the AI-powered facial recognition algorithm. As can
be seen in the figure, the page has an eye-catchy user interface, which has an image of a
playing robot as a welcoming to enter into training. On the page is shown a grid of student
photos labeled with a unique ID for each student. The system uses these labeled images to
learn and identify the faces of individual students. This page is essential to the overall
effectiveness of the system in identifying students for attendance recording.

Figure 7.5: Student Photo Training Page

Department of AI & DS, SIT, Mangaluru Page 24

Real-Time Facial Attendance System Screenshots

7.6 Student Face and Spoof Detector Page

The figure 7.6 describes Student Face and Spoof Detector Page is a critical component. It
uses advanced facial recognition to identify students accurately and catch spoofing attempts
by photograph or video instead of the real person. The page shows a live camera feed that
captures the face of the student in real time. The system then examines the captured image,
cross-checking it with the details of the enrolled student. Attendance is marked if a match
is found and the system authenticates the face as legitimate. In case of some discrepancies
or anomalies, such as the existence of a photograph or mask, the attempt will be flagged as
a probable spoofing attempt. This robust mechanism prevents unauthorized access to the
system and maintains the integrity of attendance records.

Figure 7.6: Student Face and Spoof Detector Page

7.7 Student Attendance Page

The figure 7.7 Student Attendance Page, as shown in the figure, is the central place for real-
time attendance tracking within the system. It shows a list of all students with their
attendance status, entry time, and exit time. The page gives a clear view of attendance data,
so administrators and educators can quickly track the presence of students and identify any
discrepancies. The columns present in this page are Student ID, Roll No., Student Name,
Time of Entry, Date, Attendance Status, and Exit Time. The page also offers convenient
functionalities like "Export CSV" to generate downloadable attendance reports in CSV
format for further analysis or record-keeping and "Update Exit Time" to manually adjust
exit times if needed.

Department of AI & DS, SIT, Mangaluru Page 25

Real-Time Facial Attendance System Screenshots

Figure 7.7: Student Attendance Page

7.8 Developer Page

The figure 7.8 is a contact page for the system, which is in the form of The Developer Page,
as shown below: The page shows a list of developers responsible for creating and
maintaining the system with their names. When one clicks on a developer's name, the
system will automatically take them to the developer's email address. This direct link
facilitates easy communication and allows users to quickly reach out with any queries,
doubts, or technical issues they may encounter. This feature enhances user support and
ensures that users can receive timely and effective assistance from the developers
themselves.

Figure 7.8: Developer Page

7.9 Web Camera Implemented to Classroom

The figure 7.9 describes how integration of web cameras, as shown in the image, is a vital
component of the real-time AI-Powered attendance system. The cameras are placed in
strategic locations within the classroom to record clear and consistent video footage of
students. The video feed is then processed by the facial recognition algorithm of the system
to track and monitor attendance in real time. The web cameras will support high-quality
video that should be able to offer proper resolution to identify students by names. They are
Department of AI & DS, SIT, Mangaluru Page 26
Real-Time Facial Attendance System Screenshots

also supported by features such as night vision, which helps improve performance with the
lighting conditions and even their environments. This smooth incorporation of web cameras
into classroom structures will ensure that the system captures accurate attendance
information while delivering reliable real-time insights.

Figure 7.9: Web Camera Implemented to Classroom

Department of AI & DS, SIT, Mangaluru Page 27

 CHAPTER 8

RESULT ANALYSIS
Real-Time Facial Attendance System Result Analysis

CHAPTER 8

RESULT ANALYSIS

The system consists of several advanced models designed to optimize real-time attendance
management. Figure 8.1, include a Single Face Recognition model based on AI-powered
facial recognition, which achieved an accuracy of 96% in identifying individual faces in
real time, ensuring precise and efficient attendance marking. The Multiple Faces
Recognition model had a detection accuracy of 90%, enabling the system to recognize
students accurately in small groups while maintaining consistent performance. The Group
Faces Recognition model, designed to handle larger groups with environmental
complexities, achieved an accuracy of 88%, proving its robustness under challenging
conditions.
The spoof detection mechanism enhances system security by identifying and
preventing unauthorized access attempts through masks or photographs, ensuring the
integrity of attendance data. The database integration supports functionalities like real-time
attendance tracking, exporting attendance reports in CSV format, and maintaining detailed
student profiles. The system also incorporates web camera implementation, strategically
placed to capture high-quality video feeds, including support for low-light conditions
through night vision technology.
This advanced system optimizes attendance workflows with an intuitive design,
achieving automation while maintaining accuracy and reliability. It significantly reduces
manual intervention, making it a scalable solution for organizations.

Figure 8.1: Accuracy and Validation of Real-Time AI-Powered Attendance System

Department of AI & DS, SIT, Mangaluru Page 28

 CHAPTER 9
USER MANUAL
Real-Time Facial Attendance System User Manual

CHAPTER 9

USER MANUAL

A user guide, sometimes called a user manual, is designed to assist users in using and
efficiently operating the Real-Time AI-Powered Face Recognition System with Database
Integration. The guide is a comprehensive source that explains all functionalities, features,
and processes involved in using the system, ensuring ease of use for both technical and
non-technical users.
Step 1: The starting phase is click on register or login for admin, first register with a
creating account and all details. Then login using email and password.

Figure 9.1: Admin Panel

Step 2: After login it open an attendance management panel includes all resources like
student, Face detector and attendance panels etc.

Figure 9.2: Attendance Management Panel

Step 3: Once a attendance management panel is opened then click on student panel there
admin should enter the all students details and also taking their photos.

Figure 9.3: Student Panel

Department of AI & DS, SIT, Mangaluru Page 29

Real-Time Facial Attendance System User Manual

Step 4: After the student panel work completed then move on data train panel where each
student images are train itself once you put their images in student panel details filling time
that image train here only.

Figure 9.4: Data Train Panel

Step 5: Then click on face detector panel for it start attendance taking of each student using
IP camera that implemented to class and detect using admin stored database.

Figure 9.5: Face Detector Panel

Step 6: Then next click on attendance panel where you can see attendance are stored in
CSV format of face detected of students using databases and also download option is their
for admin.

Figure 9.6: Attendance Panel

Step 7: The last panel is exit panel admin done all work then click on exit panel to close
all works.

Figure 9.7: Exit Panel

Department of AI & DS, SIT, Mangaluru Page 30

 CONCLUSION
AND
FUTURE SCOPE
Real-Time Facial Attendance System Conclusion and Future Scope

CONCLUSION AND FUTURE SCOPE

The Real-Time AI-Powered Face Recognition System, with the integration of a database,
has been an innovative and efficient solution for attendance management. By integrating
advanced facial recognition technology with user-friendly interfaces, the system ensures
accurate and real-time tracking of attendance. Key features like spoof detection,
comprehensive student details, photo training, and seamless camera integration make it a
robust platform for secure and efficient operations. Also, the ability to generate detailed
reports and its intuitive design make this system practical. This project does demonstrate
how AI technology can streamline and enhance traditional processes by providing a reliable
and scalable solution for modern attendance needs.
The system comes with many promising prospects for potential future
enhancements. Scalability can be improved to cover larger data sets and enable multiple
locations to cater to institutions and other organizations. Integrating better AI models can
enhance facial recognition capabilities, resulting in optimal working in adverse conditions.
Adding cloud support will allow integration with all other platforms, ensuring real-time
data access with safe storage. Introducing AI-driven analytics will allow for predictive
insights and trend analysis to be made in order to enhance decision-making. Expanding
compatibility with smartphones, tablets, and IoT devices will increase accessibility. Stricter
privacy measures and compliance with global data protection standards will ensure ethical
and secure usage.
The other features such as multilingual support, gamification, and notification can
be included to increase user engagement and adaptability. These developments will
position the system as a holistic and future-ready solution for attendance management. 2

Department of AI & DS, SIT, Mangaluru Page 31

REFERENCES
 REFERENCES
[1]. N. M. H. Hassan, M. A. Moussa, and M. H. M. Mahmoud, “CNN and Adaboost fusion
model for multiface recognition based automated verification system of students
attendance,” Indonesian Journal of Electrical Engineering and Computer Science, vol. 35,
no. 1, p. 133, May 2024, doi: 10.11591/ijeecs.v35.i1.pp133-139.

[2]. S. A. Feroze, S. U. R. Awan, and S. Z. Ali, “The Facial Recognition Technology in

Academic Attendance: A Comparative Study for Real-Time Management,” International
Journal of Technology Innovation and Management (IJTIM), vol. 4, no. 1, pp. 1–19, Jun.
2024, doi: 10.54489/adxn2030.

[3]. D. Balakrishnan, U. Mariappan, S. V. Kumar, M. D. Kumar, M. R. S. Reddy, and E.

M. K. Reddy, “Student Attendance Tracking Management using Face Biometric Smart
System,” IEEE, pp. 1–7, Jun. 2024, doi: 10.1109/conit61985.2024.10626516..

[4]. A. Ansari, A. Pujari, and K. Chaudhary, “Smart Attendance System: Using Facial
Recognition Technology,” International Journal for Multidisciplinary Research, vol. 6, no.
3, Jun. 2024, doi: 10.36948/ijfmr.2024.v06i03.23705.

[5]. G. A. Senthil, S. Geerthik, R. Karthikeyan, and G. Keerthana, “Face Recognition based

Automated Smart Attendance using Hybrid Machine Learning Algorithms and Computer
Vision,” IEEE, pp. 606–611, Jun. 2024, doi:10.1109/icaaic60222.2024.1074896.

[6]. T. R. Rajesh, L. N. Kuchipudi, A. Dammalapati, and S. R, “Attendance System Based

on Facial Recognition Using OpenCV,” IEEE, vol. 9, pp. 1529–1535, Jul. 2024,
doi: 10.1109/icscss60660.2024.10624932.

[7]. K. G. Oladele, E. J. Iseh, J. O. Oladele, and P. R. Bhaladhare, “AI based Face

Recognition Attendance System,” International Journal of Innovative Research in
Engineering & Multidisciplinary Physical Sciences, vol. 12, no. 2, Apr. 2024,
doi: 10.37082/ijirmps.v12.i2.230598.

[8]. A. Shankawar, R. Khan, A. Quazi, R. Sharma, and R. Gahlod, “Face Recognition

Attendance System,” Gurukul Multidisciplinary Research Journal., pp. 28–35, Jun. 2024,
doi: 10.69758/tvkw3693.

[9]. N. M. A. Thalor and N. O. S. Gaikwad, “Facial Recognition Attendance Monitoring

System using Deep Learning Techniques,” International Journal of Integrated Science and
Technology, vol. 1, no. 6, pp. 841–848, Jan. 2024, doi: 10.59890/ijist.v1i6.685.
[10]. P. Rao and N. Lakshmeesha, “Real Time Face classification and Shape based feature
extraction with Deep Neural networks,” IEEE, pp. 216–220, Dec. 2022,
doi: 10.1109/i4c57141.2022.10057731.

[11]. A. Kiran, V. Neerugatti, R. Mounika, R. Thinakaran, J. Somasekar, and N. Patidar,

“Advance attendance monitoring using facial recognition technology for seamless
tracking,” in CRC Press eBooks, 2024, pp. 465–470. doi: 10.1201/9781003559092-80.

[12]. U. M. Adeline, H. Gaspard, and K. Innocent, “A Real-Time Face Recognition

Attendance Using Machine Learning,” in Communications in computer and information
science, 2023, pp. 91–107. doi: 10.1007/978-3-031-34222-6_8.

[13]. R. K. Peddarapu, S. R. Kannareddy, B. Mallela, V. S. S. A. Prithvi, and Y. A. Reddy,

“Real Time Attendance Capturing through Face Recognition,” IEEE, pp. 726–733, Oct.
2023, doi: 10.1109/i-smac58438.2023.10290691.

[14]. N. H. Kale, N. K. Aswar, N. K. Yadav, and N. Dr. Y. Mali, “Attendance Marking

using Face Detection,” International Journal of Advanced Research in Science
Communication and Technology, pp. 417–424, Oct. 2024, doi: 10.48175/ijarsct-19961.

[15]. S. V, S. S, G. S. S, A. A, and D. R. R, “FACE RECOGNITION-BASED

ATTENDANCE SYSTEMS USING ANTI-SPOOFING METHODS,” in IIP Series, 2024,
pp. 94–112. doi: 10.58532/v3biai4p1ch6.

[16]. Z. Ning, W. Zhang, and J. Yang, “Face Anti-Spoofing Based on 3D Learnable

Convolutional Operators,” 2022 34th Chinese Control and Decision Conference (CCDC),
pp. 4034–4040, May 2024, doi: 10.1109/ccdc62350.2024.10588035.

[17]. N. Surantha and B. Sugijakko, “Lightweight face recognition-based portable

attendance system with liveness detection,” Internet of Things, vol. 25, p. 101089, Jan.
2024, doi: 10.1016/j.iot.2024.101089.