1

CHAPTER I

INTRODUCTION

Background of the Study

Face masks act as a barrier to prevent the transmission of infectious agents

between healthcare workers, patients, and visitors. This is especially important in

healthcare settings where patients may have weakened immune systems or be more

susceptible to infections. Masks help protect healthcare workers from inhaling or being

exposed to infectious droplets or airborne particles while caring for patients with

contagious illnesses. This is essential for their safety and to prevent the spread of

infections within the healthcare facility.

According to Amol Kawade, Nikita Khawase (2019), many infectious diseases,

including respiratory viruses like influenza and COVID-19, are primarily spread through

respiratory droplets when an infected person talks, coughs, or sneezes. Wearing a mask

helps contain these droplets, reducing the risk of transmission to others. In surgical

settings, masks are worn to maintain a sterile environment during procedures. They help

prevent the introduction of bacteria or other contaminants from the respiratory tract of

healthcare workers into the surgical site, reducing the risk of surgical site infections.

Overall, wearing face masks in hospitals is a fundamental infection control measure

aimed at protecting both healthcare workers and patients and preventing the spread of

infectious diseases within healthcare settings.

 2

The wearing of the face masks appears as a solution for limiting the spread of

COVID-19. In this context, the proposed project aim to create a mask detecting system

which will enable us with the information using image processing that if a person is

wearing mask on real time or not.

A face mask detection dataset consists of with mask and without mask images.

The researcher will use OpenCV to do real-time face detection. The researcher will use

the dataset to build a COVID-19 face mask detector with computer vision using Python,

OpenCV, and Tensor Flow and Keras. The goal is to identify whether the person on

image/video stream is wearing a face mask or not with the help of computer vision and

deep learning.

Statement of the Problem

Amid current pandemic, Covid-19 has made us realize the importance of Face

Masks and we need to understand the crucial effects of not wearing one, now more than

ever. Right now, there are no face mask detectors installed at the crowded places

specifically at Cruzado hospital where the proposed system will be implemented.

Accordingly, hospitals or healthcare facilities that do not strictly enforce the wearing of

face masks may encounter several problems like there is a higher risk of transmitting

infectious diseases, including respiratory viruses like influenza and COVID-19, within

the hospital environment. Healthcare workers, patients, and visitors may be more

susceptible to contracting and spreading illnesses.

This can jeopardize their health and increase absenteeism due to illness, further straining

healthcare resources. Patients who are already ill or have weakened immune systems are

particularly vulnerable to infections. Failure to wear masks can increase the risk of cross-
 3

contamination between patients and healthcare workers, leading to healthcare-associated

infections and complications. Many healthcare facilities are required to adhere to

regulations and guidelines set forth by government agencies or accrediting bodies

regarding infection control practices, including mask-wearing. Non-compliance with

these regulations can result in penalties, fines, or loss of accreditation thus the

development of the proposed system.

Objectives of the Study

Generally, the aim of the study is to develop a Face Mask Detection Application.

Specific objectives:

1. To develop a system that will make use of OpenCV for image processing;

2. To develop a system that will detect a certain person if he/she is wearing a

facemask or not using python codes; and

3. To develop a system that will integrate with a surveillance camera to enhance

safety measures.

Scope and Limitation of the Study

Face mask detection system refers to detect whether a person is wearing a mask or

not. If a person keeps on wearing a mask, then the project reads it and if a person has not

applied a mask, then reads it and tells it. And through Face Mask Detection System, the

hospital can find out which person has not applied the mask and this is my face mask

detection system security based. The Face Mask Detection System will be implemented
 4

at Cruzado Hospital. While face mask detection systems can be valuable tools for

promoting mask compliance and enhancing public health and safety, they also have

several limitations. Face mask detection systems may not always accurately identify

whether an individual is wearing a mask, especially in situations where the face is

partially obscured or the lighting conditions are poor. False positives or false negatives

can occur, leading to incorrect enforcement actions or overlooking individuals who are

not compliant. The effectiveness of face mask detection systems relies heavily on the

quality of the camera and image processing algorithms. Low-resolution cameras or poor

image quality can reduce the accuracy of mask detection, particularly in crowded or fast-

moving environments. Face mask detection systems may struggle to accurately detect all

types of masks, especially non-standard or homemade masks that may not conform to

typical shapes or designs. Additionally, the system may have difficulty distinguishing

between different types of face coverings, such as masks, scarves, or bandanas.

Significance of the Study

The study of face mask detection systems holds significant importance in several

domains, including public health, technology development, and societal well-being:

Public Health Impact: Understanding the effectiveness and limitations of face mask

detection systems contributes to efforts to control the spread of infectious diseases, such

as COVID-19. By promoting mask-wearing compliance in various settings, these systems

can help reduce transmission rates and protect public health.

 5

Technology Development: Researching face mask detection systems drives innovation

in computer vision, artificial intelligence, and image processing technologies. Studying

different approaches, algorithms, and hardware configurations can lead to the

development of more accurate, reliable, and efficient systems for detecting face masks in

real-world environments.

Cruzado Hospital: Evaluating the effectiveness of face mask detection systems helps

organizations optimize their operational processes, resource allocation, and risk

management strategies. By automating the monitoring and enforcement of mask-wearing

policies, these systems can enhance efficiency and reduce the burden on staff.

To the Future Researchers. The ideas presented may be used as reference data in

conducting new research. The outcome of the study will serve beneficial to neither

present researchers nor the future researchers for their cross-reference that will give them

background or an overview of the study.

Overall, the study of face mask detection systems is significant because it

addresses pressing public health challenges, drives technological innovation, informs

policy decisions, improves operational efficiency, promotes public awareness, and

addresses ethical and societal implications. By advancing our understanding of these

systems, we can better leverage their potential to enhance safety and well-being in

various contexts.
 6

Conceptual Framework

INPUT PROCESS OUTPUT

FACE MASK  “With Mask”

 Test dataset DETECTION  “Without Mask”
 Training dataset SYSTEM USING with beeping
 Person’s image PYTHON sound

Figure 1. The Conceptual Framework of the Study

Figure 1 shows the Conceptual Framework of the Study. Images from various

sources will be used to build a dataset. The size of datasets can be expanded by the

application of data enhancement techniques. The photographs are stored in two files,

“training dataset” and “test dataset,” each of which comprises 80 and 20% of the images,

respectively. Bounding boxes, sometimes known as “data annotations,” are created

around an area of interest using a variety of methods. Labelling pictures as “mask” or

“NO mask” will be done using the Labelling tool in the proposed system. To distinguish

between those wearing “masks” and those who aren't, the model is trained in an online

GPU environment called Google Colab. A folder referred to as “the trained folder” will
 7

be used for training purposes. Using the test folder, the model is tested for its ability to

identify and classify masks and no-masks that were found in the original photos.

Operational Definition of Terms

Data annotation is the process of labeling or tagging data to make it understandable and

usable for machine learning algorithms.

Google Colab- short for Google Colaboratory, is a cloud-based platform provided by

Google that allows users to write, execute, and share Python code collaboratively.

Face Mask- is a protective covering that is worn over the nose and mouth to help reduce

the transmission of respiratory droplets, particularly in the context of preventing the

spread of infectious diseases.

Image processing is a field of computer science and engineering that focuses on the

manipulation and analysis of digital images using various algorithms and techniques.

Keras is a Python library that is designed specifically for developing neural networks for

ML models. It can run on top of Theano and TensorFlow to train neural networks.

ML Software Library is short for machine learning library like Keras and TensorFlow.

These are libraries used in integrating Python codes.

OpenCV- (Open-Source Computer Vision Library), is an open-source computer vision

and ML software library, is utilized to differentiate and recognize faces, recognize

objects, find comparative pictures from an image database, perceive landscape and set up

markers to overlay it with

increased reality and so forth. The proposed method makes use of these features of

OpenCV in resizing and color conversion of data images.

 8

Python is a widely-used, high-level programming language known for its simplicity and

readability.

Training dataset is a collection of labeled examples used to train a machine learning

model. It consists of input data (features) and corresponding output labels or target

values.

Test dataset is a separate collection of data used to evaluate the performance of a trained

machine learning model. Unlike the training dataset, the test dataset is not used during the

training phase and is kept aside until the model has been trained on the training data.

TensorFlow is an open-source Python library specializes in what’s called differentiable

programming, meaning it can automatically compute a function’s derivatives within

high-level language.

CHAPTER II

REVIEW OF RELATED LITERATURE

Deep Learning-Based Face Mask Detection System by Khan et al. (2020)

This paper proposes a deep learning-based approach for detecting face masks in

images and videos. The authors train a convolutional neural network (CNN) using a large
 9

dataset of annotated images to classify faces as either wearing a mask or not. The system

achieves high accuracy and real-time performance, making it suitable for applications in

surveillance, public safety, and healthcare settings.

Survey on Deep Learning Techniques for Face Mask Detection by Patel et al. (2021)

This survey provides an overview of various deep learning techniques and

architectures used for face mask detection. The authors review recent research papers and

highlight the strengths and limitations of different approaches, including CNNs, transfer

learning, and object detection methods. They also discuss challenges such as dataset bias,

occlusion, and real-world deployment issues.

Real-Time Face Mask Detection System Using YOLOv3 and Deep Learning
by Nguyen et al. (2020)
This paper presents a real-time face mask detection system based on the

YOLOv3 object detection framework. The authors train the model using a custom dataset

of masked and unmasked faces and evaluate its performance on a test dataset. The system

achieves high accuracy and real-time processing speed, making it suitable for deployment

in public spaces and surveillance systems.

A Survey on Face Mask Detection and Recognition Methods During the COVID-19
Pandemic" by Wang et al. (2021)

This survey provides an overview of recent research efforts on face mask

detection and recognition methods developed in response to the COVID-19 pandemic.

The authors review approaches based on traditional computer vision techniques as well as
 10

deep learning methods. They discuss the challenges of dataset collection, model training,

and real-world deployment, and identify directions for future research.

Face Mask Detection and Recognition Using Deep Learning: A Review by Zhang et
al. (2021)

This review article summarizes recent advancements in face mask detection and

recognition using deep learning techniques. The authors provide insights into the

evolution of face mask detection systems, from early research efforts to state-of-the-art

approaches based on advanced deep learning architectures. They also discuss

applications, challenges, and future directions for research in this field.

Masked Face Recognition Using Convolutional Neural Network

Sneha Sakshi (2019) stated that recognition from faces is a popular and significant

technology in recent years. Face alterations and the presence of different masks make it

too much challenging. In the real-world, when a person is uncooperative with the systems

such as in video surveillance then masking is further common scenarios. For these masks,

current face recognition performance degrades. An abundant number of researches work

has been performed for recognizing faces under different conditions like changing pose

or illumination, degraded images, etc. Still, difficulties created by masks are usually

disregarded. The primary concern to this work is about facial masks, and especially to

enhance the recognition accuracy of different masked faces. A feasible approach has been

proposed that consists of first detecting the facial regions. The occluded face detection

problem has been approached using Multi-Task Cascaded Convolutional Neural Network

(MTCNN). Then facial features extraction is performed using the Google Face Net
 11

embedding model. And finally, the classification task has been performed by Support

Vector Machine (SVM). Experiments signify that this mentioned approach gives a

remarkable performance on masked face recognition. Besides, its performance has been

also evaluated within excessive facial masks and found attractive outcomes. Finally, a

correlative study also made here for a better understanding.

EfficientDet: Scalable and Efficient Object Detection by Tan et al. (2020)

EfficientDet is a recent advancement in object detection that focuses on achieving

a balance between model accuracy and computational efficiency. The authors propose a

scalable architecture that systematically scales up the depth, width, and resolution of the

model while optimizing for efficiency using a compound scaling method. EfficientDet

achieves state-of-the-art performance across different resource constraints, making it

highly versatile for various applications.

RetinaNet: Focal Loss for Dense Object Detection by Lin et al. (2019)

RetinaNet addresses the problem of class imbalance in object detection by

introducing a novel loss function called focal loss. This loss function assigns higher

weights to hard-to-classify examples during training, thereby focusing the model's

attention on challenging cases. RetinaNet achieves strong performance in terms of both

accuracy and robustness, particularly for detecting objects at different scales and aspect

ratios.

These studies demonstrate the growing interest and importance of face mask

detection systems, particularly in the context of public health, safety, and surveillance.
 12

Researchers are exploring various approaches, including deep learning techniques, to

develop accurate, robust, and real-time systems for detecting face masks in images and

videos.

Face Mask Detection Using Modified YOLOv4

In response to the COVID-19 pandemic, the widespread use of face masks has

become a crucial measure to reduce virus transmission. Automated system by Guo, W.,

Li, W., & Zhao, J. (2020). for detecting whether individuals are wearing masks can aid in

enforcing safety protocols in public spaces. In this paper, they propose a modified version

of the YOLOv4 object detection framework tailored specifically for face mask detection.

Our modified model achieves high accuracy and real-time performance, making it

suitable for deployment in various scenarios. We demonstrate the effectiveness of their

approach through extensive experiments on benchmark datasets and real-world scenarios.

Additionally, they discussed the integration of the proposed system into existing

surveillance systems to enhance public safety measures.

CHAPTER 3

METHODOLOGY
 13

The Prototype Model is a suitable approach for developing a face mask detection

system, especially when there is a need to gather feedback from stakeholders and validate

requirements before full-scale development. Here's how the Prototype Model could be

applied to develop a face mask detection system.

Requirements
Gathering

Quick
Refinement Design

Prototype Prototype
Testing Development

Figure 3. Phases of the Prototype Model

Work Breakdown Structure

FACE MASK DETECTION SYSTEM

USING PYTHON

Identify Design and Test Testing and

Requirements Design Prototype
Prototype Implementation

Define the Implement the Conduct testing and

Identify Scope
scope of the prototype of the evaluation of the
and objectives
prototype face mask detection prototype
system using
appropriate tools
and technologies
 14

Define the Gather

Project Schedule algorithms and feedback
techniques to
be used
Use sample datasets
Analyze or synthetic data to Validate
Requirements test the prototype Requirements
and validate its
performance
Create
Refine
Requirements
Prototype
Document

User Training

Plan for Full-

Scale
Development

Figure 2. Work Breakdown Structure

Figure 2 presents the Work Breakdown Structure for the development of the

System from the month of March, 2024 up to November, 2024.

Research Design of the Study

The research design to be used by the researcher will be descriptive method to

have a systematic investigation and study of the materials and sources of facts in order to

established and search new conclusions for the purpose of the study. Descriptive research

describes a certain present condition.

Identify Requirements
 15

Gather requirements from stakeholders, including users, administrators, and

developers. Identify the key features and functionality expected from the face mask

detection system, such as face detection, mask classification, real-time processing, and

user interface requirements.

Design Prototype

Based on the identified requirements, design a prototype of the face mask

detection system. The prototype should focus on demonstrating core functionality and

user interactions rather than implementing all features in full detail. Define the scope of

the prototype, including the types of images or videos it will process, the algorithms and

techniques it will use for face detection and mask classification, and the user interface

design.

Develop Prototype

Implement the prototype of the face mask detection system using appropriate

tools and technologies. Focus on developing the minimum requirements that showcases

the basic functionality required for detecting faces and classifying masks. Use sample

datasets or synthetic data to test the prototype and validate its performance. Iterate on the

design and implementation based on feedback from initial testing.

Test Prototype
 16

Conduct testing and evaluation of the prototype to assess its accuracy, reliability,

and usability. Test the prototype using various scenarios and input data to identify

potential issues and areas for improvement. Gather feedback from stakeholders, including

end-users, domain experts, and technical reviewers. Use feedback to refine the prototype

and address any identified issues or concerns.

Validate Requirements

Validate requirements and assumptions based on feedback from stakeholders and

testing results. Ensure that the prototype meets the needs and expectations of end-users

and aligns with the project goals and objectives. Adjust requirements and refine the

prototype as needed to address any discrepancies or gaps identified during validation.

Refine Prototype

Incorporate feedback and suggestions from stakeholders into the prototype.

Refine the user interface, improve algorithm performance, and enhance system

functionality based on the validated requirements and user preferences. Iterate on the

design and implementation of the prototype to create a more polished and robust version

of the face mask detection system.

Plan for Full-Scale Development

Once the prototype has been validated and refined, plan for full-scale

development of the face mask detection system. Use insights gained from the prototype
 17

development process to inform the development approach, resource allocation, and

project timeline. Define milestones, deliverables, and success criteria for the full-scale

development phase, and establish a roadmap for transitioning from the prototype to the

final product.

By following the Prototype Model, the researcher can create an initial version of

the face mask detection system that demonstrates core functionality, gathers feedback

from stakeholders, and validates requirements before proceeding to full-scale

development. This iterative approach helps mitigate risks, reduces development time, and

ensures that the final product meets the needs and expectations of end-users.

Flowchart of the Proposed System

Start

Input image

Detect face in an image

Does the image

use face mask?

“With Mask” “Without Mask”

With beeping sound

End
 18

Figure 4. Flowchart of the Proposed System

The flowchart for the Python programming language-based face mask detection

system is displayed in Figure 4. Image acquisition is taking a picture with a camera, then

processing the input image to improve its quality and get it ready for detection. Identify

faces using a face identification method in the pre-processed image the system will then

extract that contains each face that has been discovered. To conclude the face mask

detection procedure, the system will then save the annotated image and detection results

to a file or database for further use.

Use Case Diagram

Acquire Image

Image Processing

Image Classification/
Notification

Person Officer

Figure 5. Use Case Diagram

 19

As shown in the use case diagram above, acquiring image data from the video

camera is the first process in this system. All image data captured by the video camera

will be gathered here (which is represented with the “Acquire new image from stream

camera” process) for further data recognizing process. The next process is to detect and

recognize image data of people wearing masks or not. A dataset as a result of the learning

process is needed in the process. This dataset is obtained from the process of training the

machine to be able to recognize images of people wearing masks using training data and

testing data. The process of mask detection is represented in the actions of “Detect mask”

and checking whether the mask is detected or not. The officer will then receive a

notification sound or a beeping sound if the person detected is not wearing a mask.

Block Diagram

CLASSIFIER IMAGE PRE-

WEBCAM INPUT
DETECTS INPUT PROCESSING

IDENTIFYING IF
THE PERSON IS MACHINE
BEEP ALERT WEARING MASK LEARNING
OR NOT ARCHITECTURE

Figure 6. Block Diagram

The purpose of the proposed system is to screen person not wearing face mask.

The classifier (OpenCV) detects the input before proceeding to the pre-processing step.

The learning architecture generates result on the input image, classifying the image into
 20

mask or no mask classes. If a person is detected not wearing a mask, then a beep alert

will be generated.

Activity Diagram

Person System Officer

Acquire new image from
camera

Goes to the gate Recognize image data

Training, testing and

display notification

NO
No enter to the gate Detect?

YES

Enter to the gate

Figure 7. Activity Diagram

Figure 7 displayed the suggested system's activity diagram. The system will then

obtain a new picture from the camera of the individual who enters the gate. After the

system has recognized the image data, training and testing will be completed, and if the

individual who entered the gate was not wearing a face mask, the officer will be warned

by beeping.
 21

Respondent of the study

The respondent of the study will be the security guard and the patients of Cruzado

Hospital.

Location of the Study

The proposed study will be implemented in Cruzado Hospital, Poblacion Pikit

North Cotabato.

System Requirements

Hardware

 One (1) set of Computer Unit

o RAM: 8 Giga Bytes

o Processor: iCore 5 or Higher

o Hard Disk: One (1) terabyte or higher

o Hard Disk (Optional for Database) – One Terabyte

o Camera

Software

 Python Programming Language

 Google Colab

 Tensor Flow and Keras Machine Language

 22

REFERENCES

Guo, W., Li, W., & Zhao, J. (2020). "Face Mask Detection Using Modified YOLOv4."
2020 IEEE 7th International Conference on Cloud Computing and Intelligence
Systems (CCIS).
Khan, M. A., Khan, S., Alam, M., Alrubaian, M., & Alrubaian, M. (2020). Deep
Learning-Based Face Mask Detection System. arXiv preprint arXiv:2010.06013.

Lin, T. Y., Goyal, P., Girshick, R., He, K., & Dollár, P. (2019). Focal Loss for Dense
Object Detection. In Proceedings of the IEEE International Conference on
Computer Vision (ICCV) (pp. 2980-2988).

Mohan, A., Thiruvarangan, P., & Dhinakaran, P. (2021). "Real-time Face Mask
Detection Using Convolutional Neural Networks." 2021 International
Conference on Communication and Signal Processing (ICCSP).
 23

Nguyen, T. T., Nguyen, T. N., & Nguyen, D. C. (2020). Real-Time Face Mask Detection
System Using YOLOv3 and Deep Learning. In 2020 9th International
Conference on Computer and Communication Systems (ICCCS) (pp. 399-404).
IEEE.

Patel, A., Bhatt, N. S., & Patel, K. (2021). Survey on Deep Learning Techniques for Face
Mask Detection. In 2021 International Conference on Advances in Computing,
Communication Control and Networking (ICACCCN) (pp. 1-5). IEEE.

Tan, M., Pang, R., & Le, Q. V. (2020). EfficientDet: Scalable and Efficient Object
Detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and
Pattern Recognition (pp. 10781-10790).

Wang, C., Xu, S., Li, J., Yang, J., & Zhang, X. (2021). A Survey on Face Mask Detection
and Recognition Methods During the COVID-19 Pandemic. IEEE Access, 9,
103706-103726.

Zhang, J., Jiang, Z., Liu, J., & Xiong, L. (2021). Face Mask Detection and Recognition
Using Deep Learning: A Review. IEEE Access, 9, 73959-73974.

Zhang, Z., Yan, M., & Zhao, L. (2020). "A Novel Face Mask Detection Method Based on
YOLOv4." IEEE Access, 8, 191499-191508.