FaceAttend: Facial Recognition for Employee Attendance using Formatted: Font: (Default) Arial

Histogram of Oriented Gradient and Support Vector Machine Formatted: Indent: Left: 0 cm, First line: 0 cm, Line
spacing: single
Formatted: Left: 3.81 cm, Width: 21 cm, Height: 29.69
MARY SOL R. BALDEVARONA, NOEL P. CALISTON
cm
https://orcid.org/0000-0002-8692-0Employees021
[email protected] Formatted: Font: Bold
I Formatted: Font: (Default) Arial
Iloilo State College of Fisheriesries-; Barotac Nuevo Campus Formatted: Centered
Philippines
Formatted: Font: (Default) Arial, 10 pt
Face Attend: Employee Attendance Management System
Formatted: Font: 10 pt

MARY SOL R. BALDEVARONA Formatted: Font: 10 pt

ABSTRACT https://orcid.org/0000-0002-8692-0021 Formatted: Font: (Default) Arial, 10 pt
[email protected] Formatted: Font: (Default) Arial
Iloilo State College of Fisheries- Barotac Nuevo
Formatted: Font: (Default) Arial
Campus
Formatted: Font: (Default) Arial
Philippines
Formatted: Font: (Default) Arial
Formatted: Font: (Default) Arial, 12 pt
With the outbreak of pandemic covid19,COVID 19 pandemic, the
government has issued several means and ways to fight against it. To Formatted: Left
avoid the spreading of the f the virus and protect the employees of Formatted: Font: (Default) Arial, Bold
Institution, the researcher develops a kiosk-like touchless biometric Formatted: Centered
system that manages the employee attendance using Facial
Formatted: Font: (Default) Arial
Recognition with voice-assisted technology. The Enrollment feature of
the system will allow face image capture and register employees’ basic Formatted: Indent: Left: 0 cm, First line: 0 cm
information. The Employee and Attendance Management feature will
allow the system to manage employee information and export or email
their attendance in the form of a Daily Time Record (DTR) and recording
of attendance via time-log. Information gathered by the system is kept
in a local database. Face Detection and Recognition are performed
using Histogram of Oriented Gradient (HOG) and Support Vector
Machine Algorithm. The system was evaluated using McCall’s Software
Quality Model with a rating of 4.53 as described as “Very Effective” with
higher product quality. Therefore, the system is efficient and helpful in
managing the attendance of the employee.

Keywords: Employess Attendance Management System, Histogram of

oriented Gradient; , Facial facial Recognitionrecognition; Voice voice
aAssisted,; sSupport vVector mMachine, Philippines
With the outbreak of the COVID 19 pandemic, the Philippine government has Formatted: Font: (Default) Arial, 12 pt
issued several means and ways to fight against it. To avoid the spreading of the Formatted: Indent: Left: -0.16 cm, First line: 0 cm, Line
virus and protect the employees of the Institution, the researcher develops a spacing: single
kiosk-like touchless biometric system technology that manages employee
attendance using Facial Recognition with voice-assisted technology. The
camera will capture the image and register the employee’s basic information in
this method. The Employee and Attendance Management feature will allow the Formatted: Font: (Default) Arial, 12 pt
system to manage employee information and export or email their attendance in
the form of a Daily Time Record (DTR) and recording of attendance via time-log.
Information gathered by the system is kept in a local database. There are various
methods for comparing faces. The Histogram of Oriented Gradient(HOG) and
Support Vector Machine Algorithm is used in the computer vision problem of
face recognition. The system was evaluated using McCall’s Software Product
Quality Model in terms of Product Operation, Product Revision, and Product
Transition criteria with a rating of 4.53 as described as “Extremely useful” with
higher in Product Operation. Therefore, the system is efficient and helpful in
managing the attendance of the employee.

Keywords: histogram of oriented gradient, facial recognition, support vector

machine, Philippines, voice-assisted

*Co-authors: Noel P. Caliston

Iloilo State College of Fisheries- Dingle
Campus

INTRODUCTION

The World Health Organization(WHO) declared coronavirus disease

2019 or COVID 19 as a public health emergency of international concern last
January 30, 2020, and a pandemic on March 11, 2020. The Philippine
Government has issued several means and ways to fight against it. This virus
will come a long time until not everyone is vaccinated and following protocols.
The college always wants to protect the employee and safeguard the work
premises. The current system of recording employee attendance has various
flaws. First, attendance is not secure prone to the dishonesty of the time log,
and prone to physical contact that will be a risk to the employee and even to
the students of the ISCOF System. The current process of obtaining employee
attendance is writing on the daily time record form and not using the biometric
device first to prevent virus transmission. The result of using the manual method
is inconsistency in data entry, and the system is dependent on good individuals,
time-consuming, and lacks security. The researcher finds ways for this problem.
Along with this facial recognition system is the part in finding solutions to this
problem. The identity of any person is incomplete without facial recognition.
Just like any other form of biometric identification, face recognition requires
samples to be collected, identified, extracted with necessary (features)
information, and stored for recognition.

Facial recognition attracted huge attention due to its exemplary unique

performance and solving several complex problems. This technology gained
massive attention due to its training techniques (Qutbuddin s., et. Al, 2021). The
Face Attend is a kiosk-like standalone system that will accept an image of
employees' faces as input to record a time log. The system is also capable of
generating Daily Time Records (DTR) in PDF format for export or sending via
email, hence, it will require internet connectivity for the email function.

The framework of the Study

 HC-SR04 Ultrasonic Formatted: Font: (Default) Arial, 12 pt

Sensor
Arduino Uno R3 Formatted: Font: 10 pt
Formatted: Centered

Figure 1. Face Attend Perspective

Objectives of the Study

This study aimed to design and develop a kiosk-like touchless biometric

system in the form of Face Attend: Facial Recognition for Employee Attendance
using Histogram of Oriented Gradient and Support Vector Machine.
Specifically, this aims to: [1] develop a platform that could allow admin to enroll
employees in the system; [2] develop a platform that could allow admin to
manage employees’ records; [3] develop a platform that could allow admin to
manage attendance of employees; [4] develop a platform for the employee to
scan his face and perform attendance time log and [5] evaluate and test the
system using McCall’s software quality in terms of product operation, product
revision and product transition.

MATERIALS AND METHODS

The design part of the attendance management system is divided into two
parts: the hardware and the software. Before the software. The hardware part
is first completed to provide a platform for the software to work. Before the
software part, we need to install some libraries for the effective working of the
application. The researcher installs DlibDotNet and Visual C#.

A. Hardware Development
• Microphone for voice command
• A distance sensor is required to activate the camera and the
facial recognition system. (Ultrasonic Sensor)
• Arduino Uno R3
 • A monitor will display the time log and face recognition
identification status.
• A camera will be used to capture an image of employees.

B. Libraries Development

The system is split into two parts. [1] the creation of the front end which
is consist of GUI and backend or database of the system; and [2] the process
of attendance taking. The facial recognition of the system will be designed using
DLibDotNet developed by Takuya Takeuchi. The rest of the system will be
designed using the Windows Forms application of Visual Studio. The system
can be installed on a Windows machine only. The database is locally based.
An internet connection is required to send DTR as an attachment via email.

C. Data Acquisition

The entire face recognition solution is divided into the following major
modules: Face Capture, Face Train, and Face Recognition. Image is acquired
using a high-definition camera which is placed in the administration building of
the college. This image is given as input to the system. Dataset of employees
is created before the recognition process. Dataset was created only to train the
system. The researcher has created a dataset of five (5) faculty members,
seven (7) of which involve their names, employee numbers, email addresses,
and images of the employees from different angles and variations. For better
accuracy minimum of 10 images of each employee should be captured.
Whenever the system register employee data and images in our system to
create a dataset, employee face data is stored in the database to recall that
face in the recognition process. For face recognition the researcher used the
Histogram of oriented gradients (HOG) method is one of the feature extractors
used in computer vision and image processing. HOG Feature extraction is
primarily the characterization of the orientation and magnitude values of the pixels in
an image. That is, it defines an image in terms of groups of local histograms that point
to local regions of an image and Support Vector Machine (SVM) for classification
of the image. There are 68 specific points in a human face. The purpose of this
step is to detect landmarks of faces and to position the image. A-C # code is
used to automatically detect the face landmarks and to position the face as
much as possible without distorting the image.
 Fig. 2. Block Diagram of the System

RESULTS AND DISCUSSION

The result of the proposed system is shown below. Figure 3, The main
screen has two modes:
(1.) Normal Mode: allows employees to perform time log.
(2.) Admin mode: allows the system administrator to manage the application
 CamCCamera

Figure 3. Face Attend Main Screen

Figure 4, allows the admin to register the employee information

in the system and capture images.

Figure 4. Display Enroll Screen

Figure 5, allows the admin to manage employee records, export,
and email Daily Time Record (DTR) to the employee.
Figure 5. Display for Employee Information and Exporting DTR

Figure 6, This screen allows the administrator of the system to

manage the system settings including the time, threshold, capture
distance, port connection of the device, and save the current changes.

Figure 6. Setting Screen

Figure 7, this window shows the generated DTR of the employee.

Figure 7. Generated DTR

System Evaluation

For defining and determining the quality of a software product, the

McCall quality model contains three key perspectives: product revision (ability
to change), product transitions (adaptability to new settings), and product Formatted: Font: (Default) Arial
operations (its operation characteristics).

In terms of Product operation, Product revision, and Product transition,

the usability of the system was evaluated using McCall's software quality model.
The system was evaluated by five (5) IT Experts, five (5) Administrators, and
ten (25) employees of ISCOF. Formatted: Font: (Default) Arial

Table 1. Summary of Usability Evaluation Results

Factors IT Experts Administrators Employee Entire Group Formatted: Font: (Default) Arial, 10 pt
Formatted: Font: (Default) Arial, 10 pt
M DEC M DEC M DEC M DEC Formatted: Font: (Default) Arial, 10 pt

Product Operation Formatted: Font: (Default) Arial, 10 pt

4.20 VU 4.28 EU 4.58 EU 4.35 EU
Factor Formatted: Font: (Default) Arial, 10 pt
Product Revision Factor 4.40 EU 4.33 EU 4.47 EU 4.40 EU Formatted: Font: (Default) Arial, 10 pt
Product Transition Formatted: Line spacing: single
4.27 EU 4.13 VU 4.57 EU 4.32 EU
Factor
Formatted: Font: (Default) Arial
Overall Mean 4.29 EU 4.25 EU 4.54 EU 4.36 EU Formatted: Font: (Default) Arial
Formatted: Indent: Left: 0.16 cm, First line: 1.11 cm,
Table 1. showed the Usability summary evaluation results evaluated by Line spacing: single
the IT Experts, Administrators, and employees. It showed that the system Formatted: Font: (Default) Arial
acquired an “Extremely Useful (EU)” evaluation based on McCall’s Software
Quality Model. As a whole, all the specified criteria earned an “Extremely
Useful (EU)” in terms of Product Operation Factor (M = 4.35), Product Revision
Factor (M =4.40), and Product Transition Factor (M = 4.32). It indicates that
the system can perform a collection of operations that cover all of the given
tasks and objectives, provide accurate recognition, recording of time and
generate employee’s Daily Time records.

The overall mean of 4.36, interpreted as “Extremely Useful (EU)”,

means that the system is suitably Correct, Efficient, Reliable, Usable,
Maintainable, Flexible, and Portable.

CONCLUSIONS

• For client satisfaction, the face recognition system is made to give

accurate results. By applying accurate algorithms and quality face
recognition results can be obtained. Moreover, with the help of HOG
models and SVM, it achieves high-performance levels in recognizing
human faces and analyzing facial features, even in scenes containing
complex backgrounds.

• The facial recognition feature embedded in the attendance management

system can not only ensure attendance is taken accurately and also
eliminated the flaws in the previous system. Using technology to conquer
the defects cannot merely save resources and stop the spreading of the
virus but also reduces human intervention in the whole process by
handling all the complicated tasks to the machine.

• the system not only resolves troubles that exist in the old model but also
provides convenience to the user to access the information collected by
mailing the attendance sheet to the respected faculty.

ACKNOWLEDGMENT

The Researchers would like to thank everyone who had contributed to

the successful completion of this project. First, I would like to express my utmost
gratitude to our research Adviser, Dr. Larmie Feliztuso who despite being
extraordinarily busy with her duties, took the time to give invaluable advice and
guidance throughout the development of the research. In addition, I would also
like to express my deepest appreciation to Iloilo State College of Fisheries
Administrators, faculty, and staff for their constant support, unselfish
cooperation, and assistance that has given us to complete this task.
 LITERATURE CITED

Bekzod, et al (2021). Face recognition-based automated student attendance

system. Retrieved on September 18, 2021, from Turkish Journal of
Computer and Mathematics Education. Vol.12 No.7 (2021), 417-420

Bhatti, et al (2018. Smart Attendance Management System Using Face

Recognition. Retrieved on October 5, 2021, from EAI Endorsed
Transactions on Creative Technologies. 10 2018 | Volume 5 | Issue 17 |
e4

Howard C (2018). Face Recognition Based Automated Student Attendance

System. Retrieved on September 18, 2021, from
http://eprints.utar.edu.my/2832/1/EE-2018-1303261-1.pdf

Sakina, et al(2020). Face Recognition for Automated Attendance using HOG

and Machine Learning. Retrieved on September 16, 2021, from (PDF)
Face Recognition for Automated Attendance using HOG & Machine
Learning (researchgate.net)

Potadar, et al (2021). Attendance Management using Face Recognition.

Retrieved on September 16, 2021, from Attendance Management
system using Face Recognition (ijert.org)

Sai, et al (20210. Student Attendance Monitoring System using Face

Recognition. Retrieved on September 17, 2021, from
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3851056

Suresh, et al (2019). Facial Recognition Attendance System Using Python and

OpenCV. Retrieved on September 17, 2021, from D05021829.pdf
(questjournals.org)