CHAPTER 1

The Problem and Its Background

Introduction

Recently, a growing crisis has emerged; the

advancement of Artificial Intelligence (AI) has led to

the development of a near-quality video editing and

creation known as Deepfake. Deepfakes derive from "deep

learning" and "fake" techniques, which can superimpose

anyone's face to anyone else. These strategies categorize

into two first changing faces through images called face-

swapping or face-swap. Moreover, Lip-sync Deepfake refers

to a video that alters to create a realistic forgery of

physical features of people appearing to say or do things

that never really happened.

According to recent studies, to be able to make

Deepfake content, it needs a large number of images and

videos to train models and to create a real video and

image. Since prominent personalities like celebrities,

politicians, influencers, Etc., has so many available

videos and images online, they are typically the prime

target of Deepfake. In 2017, the first Deepfake video was

1
released across social media platforms, in which a

celebrity's face uses to swap with the face of a porn

actor.

This emerging technology was initially alarming

since image manipulation, and face-swapping create a

massive threat to violation of privacy and identity. It

holds that the advancement of Deepfake technology has the

potential to sway public opinion, mis, and

disinformation, affect election outcomes, violence, and

can even harm one person's reputation, which stole

someone's identity.

It is crucial to fight against this misleading

information and take action that tackles and identifies

Deepfake content such as fake videos and photos. As

stated by Patil & Chouragade, (2021), there are currently

no proven mechanisms for determining the authenticity of

an electronically uploaded or published digital video,

audio, or photograph. When it comes to Deepfake, superior

technology such as blockchain will come to the fore to

provide certain levels of authentication, acceptance, and

validation.

2
 With the issues and concerns regarding this matter,

the researchers propose decentralized technologies to

address this problem. Such capability is widely available

in blockchain systems. Blockchain innovation has many

uses, with the invention ready to modify and impact

different organizations, markets, and areas, including

natural food inventory networks and many others. In the

case of Deepfake, superior technology such as blockchain

will be used to provide specific levels of verification,

acceptance, and validation. Blockchain could provide

secure and tamper-proof information and transactions as a

decentralized distributed ledger. In a decentralized,

highly trustworthy, and stable manner, blockchain can

incorporate essential functionality to confirm the

authenticity and originality of digital items.

Statement of the Problem

The researchers aimed to prove and justify the

following questions:

1. What is the Impact of Blockchain Technology to Protect

the Integrity of the Video and Image to prevent

Identity Theft using Deepfake Analyzer?

3
2. How do the respondents evaluate “Deepfake analyzer

using Blockchain technology” in terms of the

following:

2.1 Functional Suitability, which was designed

based on completeness, correctness, and

appropriateness.

2.2 Performance Efficiency processes a request and

returns a response based on Time Behavior, Resource

Utilization, and Capacity.

2.3 Compatibility which system can share the same

service operations based on Co-existence and

Interoperability.

2.4 Usability based on Recognizability,

Learnability, Operability, User Error Protection,

Aesthetics, and Accessibility.

2.5 Reliability in terms of Maturity, Availability,

and Recoverability.

2.6 Security based on Confidentiality, Integrity,

and Non-Repudiation.

2.7 Maintainability based on Modifiability and

Testability.

4
Theoretical Background

Based on how blockchain technology can potentially

combat Deepfake to improve the image and video integrity.

Cheikosman et al., (2021) stated that Blockchain's

ability to provide decentralized authenticity validation

and a transparent chain of custody makes it potentially

effective as a tool to track and verify all sorts of

content. Images and videos have their cryptographic hash,

data assigned at the time of recording, or the source

origin; this is a unique fingerprint for specific content

or file. Through the study, the immutability features

that blockchain brings; the ability for a record on

blockchain to stay permanent and irreversible; and the

hash or "fingerprint" cannot be modified and becomes a

tamper-proof reference of the digital content at a

specific point in time. If applied to the Deepfake

analyzer, it means that on a blockchain network, hashes

can help prove that images, videos, or content do not

alter over time.

5
0Conceptual Framework

This study utilized the Input-Process-Output (IPO)

model in describing the overall structure of an

information processing program for the project. The

framework would be a significant guide for the readers in

understanding this study.

6
Figure 1. Input, Process, and Output

7
 Figure 1 illustrates the system's Input, Process,

and Output. The input is the users who would upload their

video or image to detect whether it is Deepfake or

natural. Then, the process would generate a hash code

based on the file that has been uploaded and then checks

it on the hash codes that have been pre-defined in the

blockchain database. When the hash code is processed, it

would give the output immediately; if not, it will start

analyzing the file. It could convert into four (4)

convolutional blocks and layers to determine each layer's

size and filter. Then, it could have a batch

normalization to improve the neural network's speed,

performance, and stability by normalizing each layer. The

tool brings numerical data to a standard scale without

distorting its shape. Then it would go to the pooling

layer, which reduces the dimensionality of the layers for

faster computation and check it to the pre-trained data

for cross-validation. Then the output would determine

whether the file is genuine or Deepfake. Then it could

display the result and automatically update the hash

database in the blockchain.

8
Objectives of the Study

This study sought to answer the following

objectives:

● to utilize decentralized technology such as a

blockchain-based approach for establishing a secure

and reliable systematic way to trace the original

video. To develop a system that will detect the

legitimacy of content.

● to analyze the impact of blockchain technology in

protecting video and image integrity from identity

theft using the Deepfake analyzer.

● to provide digital integrity across the media, which

will systematically lessen the mis and

disinformation.

Significance of the Study

This study would determine the impact of blockchain

technology on protecting video and image integrity from

identity theft using the Deepfake analyzer. Based on the

9
result that the researchers may prove and presume that

the study would contribute to the following:

To the Academicians- This study would benefit the

Institute as this project can provide extra security to

the students, staff, faculty members, office chiefs, and

administrative officials of the Institute.

To the Celebrities- The study result would also benefit

celebrities by providing them with knowledge about their

images and videos that can be used as misleading content

online called Deepfake. In addition, it will help them to

be legit- check the contents they see online.

To the Future Researchers- This study would provide

valuable information that would add and help future

researchers if they have the same study as us. This study

would serve as one of their background and knowledge

regarding blockchain technology and Deepfake. Lastly,

this study will help them improve parameters not covered

in this study.

Prominent Personalities- Given that they are widely known

personalities, we cannot stop the spread of their images

and videos online. However, using blockchain technology,

10
we can protect and avoid misleading content through our

proposed Deepfake analyzer.

To the Politicians- This study would benefit politicians

to help them secure their identity and prevent

falsification in case cybercrimes use their images and

videos online for Deepfake.

To the Society- This study would benefit society as this

would serve as security not only to high-profile

individuals but to cover everyone in the society from

possible attacks of identity theft. In addition, through

the information provided, they are less likely to fall

prey to the spread of fake news and any misleading

content online.

To the Software Developers- This study may serve as a

software program that refers to future software

developers. This research will serve as a guide and

additional information if they have the same research as

us.

Scope and Limitations of the Study

The scope of the study is the impact of blockchain

technology on protecting video and image integrity from

11
identity theft using the Deepfake analyzer. The study was

conducted in Eulogio "Amang" Rodriquez Institute of

Science and Technology (EARIST) during the First Semester

of the School Year 2021- 2022.

The study's main scope is to develop a system that

would analyze the validity of digital content. The

researcher's solution focuses only on video content, but

the solution framework provided in this paper is generic

enough and applicable to different images.

This study has the potential for different models

such as the Advanced Encryption Standard (AES) Algorithm,

Ethereum Smart Contract, and Consensus Algorithm to help

solve and automate a solution to the system. The data

sets compiled for this study are from various sources.

Furthermore, reviews of related works used in this study

were from available recent studies, news, articles, and

journals on the internet.

Lastly, the target respondents of the study were

categorized into three generations, namely Generation Z

(born from 1997 to 2012), Millennials (born from 1981 to

1996), and Generation X (taken from 1965 to 1980).

12
However, this study had several limitations—this study

limits factors like audio and other content besides

humans.

Definition of Terms

To better understand the study, the researcher gave

several terms defined conceptually and operationally as

used in the study.

AI- Artificial Intelligence

AES- It is also known as the Advanced Encryption

Standard. The United States government selected this

symmetric block cipher to safeguard sensitive data.

Analyzer- A device that identifies or examines validity.

Blockchain- Decentralized technology is used to identify

the authenticity of digital content.

Consensus Algorithm- Help determine the reliability in a

network involving multiple unreliable nodes.

Deepfake- Derives from “deep learning” and “fake”

techniques that can superimpose anyone’s face to anyone

else.

Face- swap- Changing faces through images.

13
Gen X - Generation X was born between 1965 and 1980.

Gen Z - Generation Z was born between 1997 to 2012.

Machine Learning - A machine's ability to emulate

intelligent human conduct.

Millennials - Born between 1981 to 1996

Smart Contracts- Storing programs on a blockchain that

will run when predetermined conditions meet.

14
 CHAPTER 2

Review of Related Literature

Deepfake detection is gaining popularity due to the

security risks raised by using face-swapping techniques.

This section reviews and discusses related literature and

studies on a different blockchain-based method to

determine Deepfake video content.

Foreign Studies

The study of Hasan & Salah, (2019) provided a

decentralized blockchain-based approach for protecting

and reliable traceability to the original video producer

or source of digital videos. Their method uses IPFS, a

decentralized storage system, the Ethereum name service,

and a decentralized reputation system. Ethereum

blockchain-based solutions maintain the veracity of

digital data by providing reliable and secure

traceability back to a trusted artist or publication

source. Even though the resolution concentrates on video

15
material, the solution framework, system architecture,

algorithms, sequence diagrams, and implementation and

testing details are sufficiently flexible to be used in

other forms of digital content such as audio,

photographs, images, and manuscripts.

The solution can help consumers determine if a video

or digital material can be linked to a reliable and

respectable source, which can help in the battle against

Deepfake films and audio. A movie or piece of digital

material that cannot be tracked cannot be trusted. The

smart contract-based approach provides secondary artists

with a safe means for requesting authorization from the

original artist to replicate and alter films.

Another IPFS effort Patil & Chouragade, (2021)

offered a solution, followed by the significant

components: owner, video, and IPFS. The original artist,

also known as the trustworthy owner, creates the input

video. The original artist must create an account on the

Distributed File System (IPFS). Following the completion

of registration, the first video is utilized as input.

IPFS stores the video frames and information, creating a

hash for each video. The hash value is then stored on a

16
blockchain. The technique allows social media users to

track data and assure its accuracy while gaining reliable

information from digital content.

It is another Ethereum blockchain-based ecosystem

for digital video proof of authenticity, according to

Chauhan & Kumar, (n.d.), in which safe and trustworthy

traceability to the original video creator or source may

be generated in a decentralized manner. The system

provides social media users with reliable data provenance

of digital information, allowing them to trace the data

and be sure it is correct. The approach uses Ethereum

smart contracts and the IPFS decentralized storage

technology. The IPFS discovers the files for the Ethereum

wallet, uses IPFS to store video information, and

produces a unique hash of the films. Any digital data,

such as videos and images, may be used in conjunction

with our solution structure, system design, sequence

diagrams, and implementation details. This Smart

Contract-based solution provides secondary artists with a

safe way to request permission from the original artist

to duplicate, change, share, and edit films.

17
 Tambe & Pawar, (2020) invented a technique for

proving the ethnicity of digital videos, which allows for

secure and trustworthy tracking to the underived video

producer or origin. Their method uses a Time Stamping

Technique to allow for the unique recording of critical

circumstances without endangering their actual state,

Name operation Ethereum, and a decentralized reputation

network. Their suggested system designs, algorithms,

application, and examination features are general and may

be used for various digital materials, including audio,

movies, photographs, and documents. The unravelment will

aid in detecting fake videos by assisting consumers in

determining if a movie or digital content is traced to a

reliable and reputable root.

Y. Li & Lyu, (n.d.) introduced a new deep learning-

based technique for differentiating AI-generated bogus

movies from real ones. Their method is based on the fact

that the existing Deepfake algorithm can only produce

pictures with restricted resolutions, which must be

distorted further to resemble the original faces in the

source video. They demonstrated that such modifications

result in various artifacts in the Deepfake video and

18
that convolutional neural networks (CNNs) can efficiently

capture them. In contrast to prior techniques, numerous

real and Deepfake-produced photos were used to train a

CNN classifier. Because they employed affine face warping

artifacts as the differentiating characteristic between

genuine and fake photos, their technique does not require

Deepfake-generated images as negative training samples.

Their technique has two advantages: (1) Such artifacts

may be directly reproduced on a negative example image

using simple image processing technologies because it

takes time and money to train a Deepfake model to produce

negative examples. By gathering training data, they save

time and money; (2) because such artifacts are widespread

in Deepfake movies from many sources, their approach is

more resilient than others. Their method's effectiveness

is assessed using two sets of Deepfake video datasets.

Another CNN-based technique for capturing these

semantic shift properties for Deepfake video recognition

is proposed by Ge et al., (2021). Latent Pattern Sensing,

a predictive representative learning technique, is

proposed by et al. In their technique, they utilize a

CNN-based encoder, a ConvGRU-based aggregator, and a

19
single-layer binary classifier. The encoder and

aggregator are self-trained to create the typical

spatiotemporal context characteristics. Finally, the

classifier is trained to distinguish between false and

real videos using context characteristics. Consequently,

the recovered features may coherently define the latent

patterns of films over frames, resulting in an

outstanding Deepfake video detector. Extensive testing

demonstrates the utility of our method, such as

outperforming ten state-of-the-art by at least 7.92%

@AUC.

In the Deepfake video, early efforts X. Li et al.,

(2020) presented a new challenge of partial face assault.

Although only video-level identifiers are supplied, not

all of the faces in the bogus films have been modified.

They solve this issue using multiple learning systems,

which treat faces and input video as instances and bags,

respectively. A sharp MIL (S-MIL) is presented, which

constructs direct mapping from instance embeddings to bag

prediction instead of classic MIL's mapping to instance

prediction and then to bag prediction. According to

theoretical research, the gradient vanishing in normal

20
MIL is eased in S-MIL. The spatial-temporal encoded

instance is intended to appropriately reflect intra-frame

and inter-frame inconsistency, which improves detection

performance even more. They also developed the FFPMS

dataset for partly attacked Deepfake video detection,

which can be used to evaluate different techniques at the

frame and video levels. S-MIL surpasses previous Deepfake

video detection systems for partly attacked videos in

experiments on FFPMS and the widely used DFDC dataset. S-

MIL may also be used to get best-in-class performance on

single-frame datasets in regular Deepfake image

recognition applications.

Meanwhile, Lugstein et al., (2021) concentrated on

PRNU-based detection techniques, which, while popular in

image forensics, have gotten less attention in Deepfake

detection. They employed a PRNU-based technique initially

meant to identify face morphing and retouching. They

conducted the first large-scale test of PRNU-based

Deepfake detection algorithms on several standard

datasets. We demonstrate the impact of commonly

disregarded face extraction step settings on detection

accuracy. They also demonstrate that traditional PRNU-

21
based approaches, while unable to compete with

ultramodern deep learning-based techniques, may be used

to augment them in hybrid detection systems.

The paper of Nagothu et al., (2022) proposes DeFake,

a decentralized ENF consensus-based Deepfake detection

system aimed at increasing the reliability and security

of current online conferencing systems. The PoENF

consensus mechanism is the foundation for the DeFake

system, which is resistant to byzantine nodes and

consumes low computing power. To authenticate and locate

the recordings. Each node depends on the PoENF

algorithm's ground-truth ENF. The experimental results

and evaluation validate the DeFake system's feasibility

in detecting and locating the fabrication site in

Deepfake audio streams using ENF signature

inconsistencies.

Deepfake-generating strategies were introduced and

examined in the study of Mamunur Rashid et al., (2021).

Methods for detecting deepfakes are already in use. The

research also looked at blockchain technology as a

possible solution for the negative impacts of data

manipulation and how blockchain technology might help us

22
limit the impact of harmful Deepfake technology. The

suggested framework can assist users in determining if a

video or digital material can be linked back to a

trustworthy source, which can aid in the battle against

Deepfake videos, images, and audio. They employ a more

excellent mix of revolutionary content-unique hashing

techniques, integrity methods, security measures, and

internationally adopted blockchains to build a

universally secure artificial intelligence society and

ensure that future technology is generated with good

intentions.

Local Study

Deepfake is dangerous, according to Bgen et al.,

(n.d.). They also said that a modified image or other

content might easily lead to misleading. Congress should

consider implementing legislation to address Deepfake

concerns regardless of the anticipated obstacles.

Otherwise, this technology may be used for less noble

purposes. They said that the hazards of misusing and

exploiting Deepfake technology include the capacity to

propagate fake news, influence political beliefs, and

target specific groups such as women, minorities, and

23
vulnerable individuals. Even though technology is

continuously improving, the unpleasant reality is that

the more films and photos we put publicly online, the

simpler it is to Deepfake. It can now be finished with

only one photograph of us. The most recent technology

also demonstrates how a voice-over by another actor is no

longer necessary, as text that allows new words to be

placed on someone's lips may be made.

Local Literature

The topic of Deepfake has never been new in the

Philippines. Deepfake videos are abundant on the

internet; the problem is that Filipino internet users are

unfamiliar with this technique. The viral video of a

young actor in the Philippines, Akwei, (2022), was

addressed. Andrea Brillantes' dancing TikTok video was

utilized in the Tempo App, which could face swapping the

owner's face to one's user's face. It later disturbed

numerous netizens since the problem of face-swapping may

be used as a weapon for disinformation. Specifically, it

could be used to influence Philippine elections.

24
 News, (2022) explained the impact of Deepfake and

why it is critical to educate oneself about fake news and

Deepfakes. They also offered safeguards against false

news and Deepfake. One technique is to look at the source

of the piece. The source might be counterfeit if it is

not genuine or trustworthy. Another method is to look at

the headline. If the headline looks to be too good to be

true, it most often is. Finally, enquire to see if anyone

else has heard the same thing. They also stated that

combating disinformation requires the development of

critical thinking abilities.

Synthesis

All integration-related works summarize that

Deepfake is a looming technological threat that needs

prevention. Since Deepfakes continue to rise, researchers

have developed and proposed different blockchain models

to detect fake videos and images. There are several

methods for working with detection, recognition, and

analyzers. Based on the analyzed articles, most related

works used a decentralized blockchain platform such as

the Ethereum blockchain approach since it provides a path

forward for ensuring data integrity from data

25
manipulation. Based on the reviewed materials and the

Ethereum blockchain, Smart Contracts, and Interplanetary

File systems are the most used and effective methods to

apply and help protect data integrity.

Given the review materials, neural networks like

Convolutional Neural Networks and Generative Adversarial

Networks have been viewed as one of the most popular

techniques for identifying Deepfake because they reduce

errors, which lowers the difference between the fake

image and video content and the actual image and video

content.

Based on the review of related works, the presented

models are helpful to our proposed research as they will

help us develop and guide our study. As a result, the

related literature and studies have one fundamental

relationship with current research that is also focused

on developing a Deepfake analyzer using blockchain

technology.

26
 CHAPTER 3

Research Methodology

This chapter presents the methodology used to

conduct the research study. This study advocates several

ways in which the use of different research methods can

be truly complementary. It covers the topic such as the

method of the research, subjects respondents of the

study, research locale, sampling procedure, instrument

and techniques used in conducting the study, the process

of data gathering, and data analysis.

The researchers used a descriptive-quantitative

research approach to obtain more specific and reliable

data from the study’s respondents and participants. A

descriptive approach is used because it provides a

sufficient and precise interpretation of more extensive

27
sample data. Overall, the research method chosen will be

a valuable tool for analyzing and describing the survey-

questionnaire data.

Requirement Analysis

28
 Figure 2 System Flowchart

In Figure 2, the researchers developed a system

called "Deepfake Analyzer," which will enable it to track

and verify images or videos using a hash to help the

researchers and users analyze the impact of blockchain on

the integrity of images or video content found online.

Population, Sample Size, and Sampling Techniques

The purposive sampling method was used to develop

the research sample under discussion. According to this

method, which belongs to the non-probability sampling

techniques, sample members are of their knowledge,

relationships, and expertise regarding a research subject

(Freedman et al., 2007). In the current study, the

selected sample members had a special relationship with

the phenomenon under investigation, sufficient and

relevant interest, or experience in the field of

Technology. A total of ninety respondents will be

29
purposively selected as participants in a pre-survey. The

objective of the pre-survey is to provide information

about the respondents that can be used later to determine

whether these respondents have changed concerning the

factors measured by the survey. To measure the amount of

change, a post-survey that measures the same elements is

administered to the subject population. Purposive

sampling is a non-probability sample selected based on

the population's characteristics and the study's

objective.

Respondents of the Study

Three generations—Generation Z (born between 1997

and 2012), Millennials (born between 1981 and 1996), and

Generation X—will be chosen as respondents for this study

(born from 1965 to 1980). For the pretesting stage of the

survey research, the researcher will select thirty

respondents from Generation Z, thirty from Generation X,

and thirty from Millennials. The goal of the pretest is

to evaluate the reliability and validity of the survey

instruments before the final distribution. For the post-

survey of the research, the researcher will select ten

30
experts in the field of Technology and research

development.

Table 1

Respondents of the study in pre-survey in terms of their

Type

Type of respondents Sample

Generation X 30

Generation Z 30

Millennial 30

Total 90

Table 1 shows the selected respondents were chosen

based on their efficiency to the researchers because it

aided them more in the testing phase of the developed

system and the surveys.

31
Table 2

Distribution of type of respondents in pre-survey in

terms of their sex

Generation Millennial Row Total

Generation Y
X

f % f % f % f %

Male 6 20.0 13 43.3 9 30.0 28 31.1

Female 24 80.0 17 56.7 21 70.0 62 68.9

Colum Total 30 100. 100.0 100. 100

30 30 90
0 0

where: f – frequency, %- percentage

Table 2 shows the profile of respondents in terms of

their sex. It reveals that 31.1% are male and 68.9% are

female. Thus, the majority of the respondents in the pre-

survey are female.

32
Table 3

Assessment of Generation X, Generation Z, and Millennial

in the pre-survey questions

Generation Generation Millennial Overall

Pre-survey
X Y
Questions
WM VD WM VD WM VD WM VD

1. Do you agree 4.80 SA 4.53 SA 4.80 SA 4.71 SA

that it is

important to

fight against

false

information and

to take an

action about

“Deepfake”?

33
2. Do you agree 4.73 SA 4.43 A 4.63 SA 4.60 SA

that “Deepfake”

is a

significant

threat to

society

affecting

people’s

perception and

misleading

them?

3. Do you agree 4.90 SA 4.63 SA 4.57 SA 4.70 SA

that “Deepfake”

can be used to

spread fake

news and lead

to an increase

in fake

scandals?

4. Do you agree 4.70 SA 4.33 A 4.43 A 4.49 A

that “Deepfake”

technology have

the potential

to sway public

opinion?

34
5. Do you agree 4.87 SA 4.60 SA 4.50 SA 4.66 SA

that “Deepfake”

may cause more

severe problems

such as

violation of

the human

rights, right

of privacy,

personal data

protection

right, etc.

6. Do you agree 4.73 SA 4.33 A 4.30 A 4.46 A

that “Deepfake”

can potentially

distort

democratic

discourse such

as manipulate

elections and

political

views?

7. Do you agree 4.90 SA 4.50 SA 4.53 SA 4.64 SA

that “Deepfake”

35
 can be misused

to commit

fraud,

extortion,

bullying as

well as to

falsify

evidence?

8. Do you agree 4.77 SA 4.47 A 4.50 SA 4.58 SA

that social

media platforms

like Facebook,

Google, or

YouTube plays

an important

role in

alerting the

public about

faked images or

videos?

9. Do you believe 4.77 SA 4.60 SA 4.70 SA 4.69 SA

that it is

important to

teach society

36
 knowledge about

“Hoaxes and

Fake Content”

to help them

learn more

about

misinformation,

and

disinformation,

online?

10. Do you agree 4.60 SA 4.40 A 4.57 SA 4.52 SA

that it is

important to

build a system

utilizing

“Blockchain”

Technology to

combat

“Deepfake”?

Overall Mean 4.78 SA 4.48 A 4.55 SA 4.60 SA

*WM-Weighted Mean; VD-Verbal Description

Legend

4.50-5.00 SA Strongly Agee

3.50-4.49 A Agree

2.50-3.49 MA Minimally Agree

1.50-2.49 D Disagree

37
 1.00-1.49 SDA Strongly Disagree

Table 3 presents the assessment of Generation X,

Generation Z, and Millennials in the pre-survey

questions, and they strongly agree on the following

questions: “Do you agree that it is important to fight

against false information and to take an action about

Deepfake?” (WM = 4.71), “Do you agree that Deepfake

can be used to spread fake news and lead to an

increase in fake scandals?” (WM=4.70), “Do you believe

that it is important to teach society knowledge about

Hoaxes and Fake Content to help them learn more about

misinformation, and disinformation, online?”

(WM=4.69), “Do you agree that “Deepfake” can be

misused to commit fraud, extortion, bullying as well

as to falsify evidence?” (WM=4.64), and “Do you agree

that Deepfake is a significant threat to society

affecting people’s perception and misleading them?”

(WM=4.60) Furthermore, they also strongly agree in “Do

you agree that social media platforms like Facebook,

Google, or YouTube plays an important role in alerting

the public about faked Images or videos?” and “Do you

38
 agree that it is important to build a system utilizing

“Blockchain” Technology to combat “Deepfake”?”

(WM=4.52). Likewise, the three types of respondents

agree in “Do you agree that it is important to build a

system utilizing “Blockchain” Technology to combat

“Deepfake”? (WM = 4.60), “Do you agree that “Deepfake”

technology has the potential to sway public opinion?”

(WM=4.49), and “Do you agree that “Deepfake” can

potentially distort democratic discourse such as

manipulate elections and political views?” (WM =

4.46).

Overall, the three types of respondents strongly

agree with this set of questions. This pre-survey

implies that there is indeed a problem on Deepfake

that needs to take care of and resolved.

Data Gathering Instruments and Techniques

For this project study, data were be gathered. The

need for constructing the suggested system was recognized

after extensive analysis and data collection. Data

gathering is essential since it shows how to apply the

proposed method properly, accurately, and effectively.

39
 The researchers used Survey Questionnaire,

Observation, and Existing Data as data-gathering

instruments.

 Questionnaire- The questionnaire is one of the

instruments used to process and collect data from

respondents, and it consists of a series of

questions. The survey is critical to the current

research since it will collect information essential

to designing the appliance. By using it, researchers

will be able to acquire much knowledge in a much

shorter period. Furthermore, using the survey for

data collection, there is less chance of bias and

impartiality in the data collected.

In this study, researchers used Google Form to

distribute the survey surveys. Google Forms allowed

the researchers to collect data in a variety of ways.

A form can be inserted in the body of an email that

will enable respondents to submit comments directly

from their Inbox. A link can be created that allows

40
 responders to respond to the questions using a web-

based form.

 Existing Data- The secondary quantitative research

approach employs secondary data or data that has

already been collected. Existing Data is summarized

and collated to improve the overall effectiveness of

the study. Conducting research studies on existing

systems and applications to identify the processes and

features that will be improved as the new application

is developed.

Data Gathering Procedure

To conduct the study, a letter of request to

disseminate questionnaires was presented to the research

adviser and statistician. Upon approval, questionnaires

were personally distributed by the researchers to the

selected respondents, namely Generation Z (born from 1997

to 2012), Millennials (born from 1981 to 1996), and

Generation X (taken from 1965 to 1980).

The researchers also presented a letter of request

to the respondents to voluntarily choose to participate

or not and answer the survey questionnaire. The

41
researchers oriented the respondents who would answer the

survey. Informants can choose their preferred response in

the study. Moreover, to assure respondents' cooperation

in this survey, the researchers guaranteed that the

information gathered would be treated confidentially. It

will be noted if responders refuse to give part of the

highly secret personal information. The data was

collected right after the selected respondent answered

the questionnaire. Afterward, the researcher will collect

the questionnaires form from the respondent and start

interpreting the data.

Preparation of Instrument

Survey-questionnaire was used to collect the data

for this study. This instrument was chosen because it

allows researchers to efficiently get all completed

replies from respondents. Respondents provide data, which

researchers use to evaluate current processes and make

recommendations about enhancing the research process.

This is more effective and can potentially reach a

considerable number of individuals simultaneously. In

addition, to answer concerns regarding the usage of the

current system, the researchers carried out the survey

42
utilizing the system’s standards and criteria of ISO /IEC

25010:2011, including functionality, reliability,

usability, efficiency, maintainability, and portability.

The questionnaire was designed by the researchers

called "Deepfake Technology: Impact of Blockchain

Technology to Protect Video and Image Integrity from

Identity Theft Using Deepfake Analyzer "in this study.

The instrument's content was based on the information

gathered during the literature review to know the impact

of blockchain technology on protecting video and image

integrity from identity theft using the Deepfake

analyzer.

The questionnaire was divided into two (2) sections

and three (3) sets of questions:

I. Section 1: Data Privacy Consent

1. Section 2: Set of Questions

a) Set 1: Respondents Profile

b) Set 2: Likert Scale

Statistical Treatment of Data

43
 The following statistical tools were used to

interpret the results according to sub-problems. In the

study, the following statistical methods were used:

1. Frequency- The respondent’s response to a particular

item/question in the questionnaire, where he or she

checks their option.

2. Weighted Mean / Likert Scale- Right after the data

collected from the field were analyzed.

Statistically, the weighted mean was used in

answering the research questions. The response

options in the instrument are weighted as shown

below:

Table 4. Likert’s Scale to evaluate the result and its

Descriptive Interpretation

INTERVAL RATING INTERPRETATION

4.20-5.00 5 Strongly Agree

(SA)

3.40-4.19 4 Agree (A)

2.60-3.39 3 Minimal Agree

(MA)

1.8-2.59 2 Disagree (D)

44
 1.00-1.79 1 Strongly Disagree

(SD)

Table 4 shows the descriptive interpretation of

evaluation on the Likert Scale. The Likert scale was

interpreted as “5” strongly agree, “4” as agree, “3”

minimally agree, “2” as disagree, and “1” as strongly

disagree.

3. Ranking- This was used to reinforce the percentage

to show the proportional importance of an idea

considered.

4. Percentage- Used to describe descriptive statistics

or parts of a whole.

f
%f = ×100
n

Where:

%ƒ – relative frequency

ƒ- frequency

n - number of observations or sample size.

5. Weighted Mean- It was used to represent the

respondents’ midpoint answer of the qualitative

45
 response. This was a criterion for determining the

status.

The weighted mean was used as follows:

∑ wx
wx=
∑w

Where:

wx - weighted mean

x - any particular value

n - number of observation or sample

Σ - summation symbol means to “sum up”

6. Independent Sample Testing- An independent sample

test is a statistical test that determines if two

groups differ significantly in terms of a variable

of interest. The researchers want to investigate the

gap between user ratings and expert ratings for each

system function.

x 1−x 2
t=

√
2 2
s 1 s2
+
n1 n2

Where:

46
 t - t value

x 1−mean of the first group

x 2−mean of the second group

2
s1−variance of the first group

2
s2−variance of the first group squared

n1 −sample ¿the first group

n2 −sample ¿the second group

Analytical Tool

Analytical tools are representations of a structure

in graphical or visual form. These tools are essential

for the researchers because they act as communication and

planning tools, identify roles, provide an overview of

the proposed system, promote logical procedures,

demonstrate relationships, aid in troubleshooting, and

document the system for future use. The researchers used

analytical techniques to give readers a rapid overview of

the suggested system design. The various graphical models

depict the system’s boundaries and the information used

in the proposed system. These tools gave the researchers

47
a better understanding of the existing and developed

system.

The following diagrams adequately described and give

an insight into the study project that the researchers

proposed; the first is the system flowchart, which shows

the flow of the data in the system and how decisions are

made to control events. The second is the context

diagram, which presents the external entities with whom

the system interacts. Next is the data flow diagram,

which presents the accurate flow of data through a

process or a system. Then the use case diagram, this

graphical illustration, shows the possible interaction

between the end-user and the system. The entity-

relationship diagram describes the interrelated things of

interest in a specific knowledge domain.

Software Model

48
 Figure 3 The Scrum Model

Figure 3 illustrates the type of method used in "The

Impact of Blockchain Technology to Protect Video and

Image Integrity from Identity Theft using Deepfake

Analyzer" is the Scrum method. The process involves

iterative practices that lean towards a defined goal or

objective. In planning the developed system, three

leading roles were involved in the development process:

the project manager, the scrum master, and the scrum

team. The project manager is the member responsible for

giving inputs to ensure all expected outcomes will be in

the final output. In terms of scrum master, this member

facilitates the scrum team and is responsible for

49
overseeing that the scrum process and practices are

followed to ensure the project is moving forward.

Regarding a scrum team, the researchers worked

towards the main tasks in the scrum, which consisted of

development, analysis, and testing. The scrum methodology

integrated a framework that facilitated team

collaboration through its 5 phases product backlog

creation, sprint planning and sprint backlog creation,

the sprint and daily scrum meetings, the sprint review,

and the sprint retrospective. Regarding product backlog,

the researchers brainstormed what features and

requirements would be implemented during the development

process. In terms of sprint planning and sprint backlog,

the researchers determined the sprint's duration and the

goal or tasks that were assigned for each sprint that the

scrum team is capable of finishing on time.

Regarding sprint and daily scrum meetings, the

researchers started the development process according to

the project's specifics and conducted daily scrum

meetings that provide the current project status,

ensuring that all team members are on track. Regarding

the sprint review, the researchers analyzed the

50
development progress and overall results for planning the

next sprint and further project changes. The results were

discussed in terms of sprint retrospectives to find ways

to improve the scrum team's development process for

future sprint iterations.

System Requirements

The researchers utilized a set of system

requirements to be able to accomplish the study project.

This requirement aims to provide a better-quality system

that will be presented to the target respondents and the

academe.

The system requirements that the researchers provide

were a set of minimum specifications that the researchers

used in the process of developing the system project. In

line with this, the researchers sought to provide.

Hardware

The researchers will utilize this set of hardware

devices to create, build, and assess the system's

quality.

51
OS: Windows 11 Pro

Version 21H2

OS build 22000.739

Experience Windows Feature Experience Pack

1000.22000.739.0

CPU: Processor AMD Ryzen 3 2200G with Radeon Vega

Graphics 3.50 GHz

GPU: Nvidia 1660 Ti (6GB)

Ram: Installed RAM 16.0 GB (2 x 8 2666 mhz)

System type: 64-bit operating system, x64-based processor

Disk Storage: 4 GB of free disk space

Monitor Resolution: 1920x1080

Software

The researchers used a specific Integrated

Development Environment (IDE) to develop the research

system. The different libraries were also utilized to

develop and enhance the accuracy of the research system.

IDE: Anaconda Integrated Development Environment version

2.2.0 Anaconda IDE has simplified package management

52
consisting of different Python and R Programming

languages for scientific computing.

Jupyter Notebook was utilized in developing the system

project; it is an open-source web application that the

end-users can use to create and share documents

containing live codes, equations, visualization, and

text.

Python 3.5 -

Python is a high-level, interpreted, general-purpose

programming language. Its design philosophy emphasizes

code readability with the use of significant indentation.

Numpy 1.14.2 - NumPy is the fundamental package needed

for scientific computing with Python.

Keras 2.1.5 - Deep Learning for Python

imageio 2.19.3 - A Python library for reading and writing

image data.

ffmpeg 5.0.1 - Cross-platform solution to record, convert

and stream audio and video.

face_recognition 1.3.0 - Recognize faces from Python or

from the command line.

53
tensorflow 2.8.1 - TensorFlow is an end-to-end open-

source platform for machine learning

Peopleware

The following experts provided a better

understanding of the topic of this study and gave an

insight into the process of creating the system. Firstly,

there is the data scientist, whose specialty is

analyzing, processing, and modeling data before

interpreting the findings to provide practical business

and other organizational strategies. Second are the

digital forensics specialists with backgrounds in

criminal justice and cybersecurity. Next are the experts

in machine learning who specialize in creating

algorithms. After that, the cybercrime division

implements cybercrime prevention and suppression laws,

performs vulnerability assessments, and aids various law

enforcement agencies with technical investigative

support. Next are Computer Science and Information

Technology professors, whose expertise is in teaching a

wide range of technical knowledge. Finally, the system's

eventual users will take advantage of the research team's

creation.

54
Network

Figure 4 The Network

Data Warehouse

The researchers used a database for the pre-trained

Deepfake datasets for image and video cross-validation. A

blockchain database also stores the hash of the processed

images and video.

CHAPTER 4

Presentation, Analysis, and Interpretation of Data

55
 This chapter describes the analysis of data followed

by a discussion of the research findings. The findings

relate to the research questions that guided the study.

Data were analyzed to identify, describe, and explore the

Impact of blockchain technology to protect video and

image integrity from identity theft using a deepfake

analyzer. Data were obtained from self-administered

questionnaires validated by professionals such as

researchers Statisticians and Thesis Advisers. The

researchers used two research participants; (1) research

beneficiaries and (2) experts related to the study. This

study was evaluated in terms of International

Standardization Organization (ISO/ IEC) 25010:2011 and

functional sustainability, performance efficiency,

compatibility, usability, reliability, security based on

confidentiality, integrity, and non-repudiation, and

maintainability based on modifiability and testability.

Table 5
Assessment of respondents in Functional Suitability

Indicators Expert User Composite

Mean
WM VI WM VI WM VI

56
The system covers all the 4.40 A 4.37 A 4.37 A
specified tasks and user
objectives.
(Functional Completeness)

The system provides the 4.40 A 4.26 A 4.27 A

correct results with the
needed degree of
precision.
(Functional Correctness)

The system facilitates the 4.70 SA 4.11 A 4.17 A

accomplishment of
specified tasks and
objectives.
(Functional
Appropriateness)

Overall Weighted Mean 4.49 A 4.24 A 4.27 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean
Point Range Verbal Symbol
Interpretation
5 4.50- Strongly Agree SA
5.0
4 3.50- Agree A
4.49
3 2.50- Minimally Agree MA
3.49
2 1.50- Disagree D
2.49
1 1.00- Strongly SDA
1.49 Disagree

Table 5 demonstrates the assessment of respondents

in Functional Suitability. It was shown that the

respondents agreed that the system covers all the

specified tasks and user objectives Functional

Completeness (WM=4.37). Furthermore, the respondents

agreed that the system provides the correct results with

the needed degree of precision Functional Correctness

(WM=4.27). On the other hand, the respondents agreed that

57
the system facilitates the accomplishment of specified

tasks and objectives of Functional Appropriateness

(WM=4.17).

Overall, the respondents strongly agreed with the

Functional Suitability of the developed Impact of

blockchain technology to protect video and image

integrity from identity theft using a deep fake analyzer

system (OWM=4.27).

Table 6
Assessment of respondents in Performance Efficiency

Indicators Expert User Composite

Mean
WM VI WM VI WM VI
The system’s response and 4.80 SA 4.20 A 4.26 A
processing times and
throughput rates when
performing its functons,
meet requirements.
(Time Behavior)

The system’s amounts and 4.70 SA 4.04 A 4.11 A

types of resources used
when performing its
functions, meet
requirements.
(Resource Utilization)

The system’s maximum 4.50 SA 4.10 A 4.14 A

limits of parameter meet
requirements.
(Capacity)

Overall Weighted Mean 4.67 SA 4.11 A 4.17 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean
Table 6 demonstrates the assessment of respondents

in Performance Efficiency. It was shown that the

58
respondents agreed that the system response and

processing times and throughput rates when performing its

functions meet the requirements Time behavior (WM=4.26).

Furthermore, the respondents agreed that the system

amounts and types of resources used when performing its

functions meet the requirements Resource Utilization

(WM=4.11). On the other hand, the respondents agreed that

the system’s maximum limits of parameters meet the

requirements Capacity (WM=4.14).

Overall, the respondents strongly agreed with the

performance efficiency of the developed Impact of

blockchain technology to protect video and image

integrity from identity theft using a deep fake analyzer

system (OWM=4.17).

Table 7
Assessment of respondents in Compatibility

59
 Indicators Expert User Composite
Mean
WM VI WM VI WM VI
The system can perform its 4.50 SA 4.24 A 4.27 A
required functions
efficiently while sharing
a common environment and
resources with other
products, without
detrimental impact on any
other product.
(Co-existence)

The system can exchange 4.40 SA 4.14 A 4.17 A

information and use the
information that has been
exchanged.
(Interoperability)

Overall Weighted Mean 4.45 SA 4.19 A 4.22 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 7 demonstrates the assessment of respondents

in Compatibility. It was shown that the respondents

agreed that the system can perform its required functions

efficiently while sharing a common environment and

resources with other products, without detrimental impact

on any other product Co-existence (WM=4.27). Furthermore,

the respondents agreed that the system can exchange

information and use the information that has been

exchanged Interoperability (WM=4.17).

Overall, the respondents strongly agreed with the

compatibility of the developed Impact of blockchain

technology to protect video and image integrity from

60
identity theft using a deep fake analyzer system

(OWM=4.22).

Table 8
Assessment of respondents in Usability

Indicators Expert User Composite

Mean
WM VI WM VI WM VI
The system allows users to 4.60 SA 4.02 A 4.08 A
recognize if it is
appropriate for their
needs.
(Appropriateness
Recognizability)

The system can be used by 4.50 SA 4.04 A 4.09 A

specified users to achieve
specified goals of
learning to use the
application with
effectiveness, efficiency,
freedom from risk and
satisfaction in a
specified context of use.
(Learnability)

The system has attributes 4.60 SA 4.12 A 4.17 A

that make it easy to
operate and control.
(Operability)

The system protects users 4.40 A 4.08 A 4.11 A

against making errors.
(User Error Protection)

The system’s user 4.30 A 4.17 A 4.18 A

interface enables pleasing
and satisfying interaction
for the user.
(User Interaction
Aesthetics)
The system can be used by 4.70 SA 4.14 A 4.20 A
people with the widest
range of characteristics
and capabilities to
achieve a specified goal
in a specified context of
use.

61
(Accessibility)

Overall Weighted Mean 4.52 SA 4.10 A 4.14 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 8 demonstrates the assessment of respondents

in Usability. It was shown that the respondents agreed

that the system allows users to recognize if it is

appropriate for their needs Appropriateness

Recognizability (WM=4.08). Furthermore, the respondents

agreed that the system can be used by specified users to

achieve specified goals of learning to use the

application with effectiveness, efficiency, freedom from

risk and satisfaction in a specified context of use

Learnability (WM=4.09). In addition, the respondents

agreed that the system has attributes that make it easy

to operate and control Operability (WM=4.17). On the

other hand, the respondents agreed that the system

protects users against making errors User Error

Protection (WM=4.11). Moreover, the respondents agreed

that the system user interface enables pleasing and

satisfying interaction for the user, User Interaction

Aesthetics (WM=4.18). Along with this, the respondents

agreed that the system can be used by people with the

62
widest range of characteristics and capabilities to

achieve a specified goal in a specified context of use

Accessibility (WM=4.20).

Overall, the respondents strongly agreed with the

usability of the developed Impact of blockchain

technology to protect video and image integrity from

identity theft using a deep fake analyzer system

(OWM=4.14).

Table 9
Assessment of respondents in Reliability

Indicators Expert User Composite

Mean
WM VI WM VI WM VI
The system meets the needs 4.50 SA 4.19 A 4.22 A
for reliability under
normal operation.
(Maturity)

The system is operational 4.50 SA 4.11 A 4.15 A

and accessible when
required for use.
(Availability)

The system can recover the 4.50 SA 4.07 A 4.11 A

data directly affected and
re-establish the desired
state.
(Recoverability)

Overall Weighted Mean 4.50 SA 4.12 A 4.16 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 9 demonstrates the assessment of respondents in

Reliability. It was shown that the respondents agreed

that the system meets the needs for reliability under

63
normal operation Maturity (WM=4.22) Furthermore, the

respondents agreed that the system is operational and

accessible when required for use Availability (WM=4.15).

In addition, the respondents agreed that the system is

can recover the data directly affected and re-establish

the desired state Recoverability (WM=4.11).

Overall, the respondents strongly agreed with the

reliability of the developed Impact of blockchain

technology to protect video and image integrity from

identity theft using a deep fake analyzer system

(OWM=4.16).

Table 10
Assessment of respondents in Security

Indicators Expert User Composite

Mean
WM VI WM VI WM VI
The system ensures that 4.50 SA 4.18 A 4.21 A
data are accessible only
to those authorized to
have access.
(Confidentiality)

The system prevents 4.50 SA 4.14 A 4.18 A

unauthorized access to,
or modification of,
computer programs or
data.
(Integrity)

The system can be proven 4.30 A 4.19 A 4.20 A

to have taken place, so
that the events or
actions cannot be
repudiated later.
(Non-repudiation)

64
 Overall Weighted Mean 4.43 A 4.17 A 4.20 A
Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 10 demonstrates the assessment of respondents

in Security. It was shown that the respondents agreed

that the system ensures that data are accessible only to

those authorized to have access Confidentiality (WM=

4.21). Furthermore, the respondents agreed that the

system prevents unauthorized access to, or modification

of, computer programs or data Integrity (WM= 4.18). On

the other hand, the respondents agreed that the system

can be proven to have taken place, so that the events or

actions cannot be repudiated later Non-repudiation

(WM=4.20).

Overall, the respondents strongly agreed with the

security of the developed Impact of blockchain technology

to protect video and image integrity from identity theft

using a deep fake analyzer system (OWM=4.20).

Table 11
Assessment of respondents in Maintainability

65
 Indicators Expert User Composite
Mean
WM VI WM VI WM VI
Reducing inter-service 4.50 SA 4.19 A 4.22 A
dependencies enhances the
system's modifiability.
(Modifiability)

The system can be tested 4.50 SA 4.22 A 4.25 A

using a functional
testing instrument.
(Testability)

Overall Weighted Mean 4.50 SA 4.21 A 4.24 A

Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 11 demonstrates the assessment of respondents

in Maintainability. It was shown that the respondents

agreed that the system reduced inter-service dependencies

enhances the system's modifiability (WM=4.22).

Furthermore, the respondents agreed that the system can

be tested using a functional testing instrument

Testability (WM=4.25).

Overall, the respondents strongly agreed with the

maintainability of the developed Impact of blockchain

technology to protect video and image integrity from

identity theft using a deep fake analyzer system

(OWM=4.24).

Table 12
Summary of assessment of respondents in enumerated system
features

66
 Weighted Verbal Rank
Indicators Mean Interpretation
Functional Suitability 4.27 Agree 1
Performance Efficiency 4.17 Agree 5
Compatibility 4.22 Agree 3
Usability 4.14 Agree 7
Reliability 4.16 Agree 6
Security 4.20 Agree 4
Maintainability 4.24 Agree 2
Note: WM - Weighted Mean, VI – Verbal Interpretation,
OWM – Overall Weighted Mean

Table 12 summarizes the assessment of respondents in

different System Features. It was exposed that the

assessment in functional suitability got the highest

weighted mean of 4.68 and was interpreted as agree. This

because the respondents determines that the systems

provide all the functions that meet all the stated or

implicit needs when used under specific conditions. On

the contrary, the assessment in usability got the lowest

weighted mean of 4.47 and interpreted as agree. Given the

fact that the system is user- friendly, some of the

respondents might not learn it easily. Due to this the

research come up with the idea of focusing on improving

the user usability in a way of, making the system ease-

of- use during the design process and user experience.

Table 13

67
Post-analysis of the Differences between the assessment
of IT Experts and Customers in the Enumerated Features of
the Impact of blockchain technology to protect video and
image integrity from identity theft using deep fake
analyzer.

Indicators t-value p-value Decision Interpretation

Functional 1.826 0.071 Fail to No
Suitability Reject Ho Significance
Performance 3.096 0.003 Reject Ho Significant
Efficiency
Compatibility 1.384 0.170 Fail to No
Reject Ho Significance
Usability 2.608 0.011 Reject Ho Significant
Reliability 2.157 0.033 Reject Ho Significant
Security 1.402 0.164 Fail to No
Reject Ho Significance
Maintainability 1.700 0.192 Fail to No
Reject Ho Significance
t-ratio α = 0.05; df = 98; Critical value = 1.984
If Computed t-value > Critical value, then reject Ho.

Table 13 displays the post-analysis of the

differences in the assessment of Experts and Users in the

enumerated features of the Impact of blockchain

technology to protect video and image integrity from

identity theft using a deep fake analyzer. It was

revealed that the associated computed t-values for

functional suitability (t=1.826), compatibility

(t=1.384), security (t=1.402), and portability (t=1.700)

are all less than 1.984 critical value (df=98, α = 0.05)

of This implies that there is not enough evidence to

reject the null hypothesis and conclude that there is no

significant difference between the assessment of experts

68
and users in the abovementioned enumerated features. The

perception of both respondents in the enumerated features

is excellent and it was tested that there is no

difference between them. Therefore, both respondents

accepted the Impact of blockchain technology to protect

video and image integrity from identity theft using a

deepfake analyzer in their view.

Conversely, the associated t-values for performance

efficiency (t=3.096), usability (t=2.608), and

reliability (t=2.157) are all greater than 1.984 critical

value (df=98, α = 0.05). This yields the rejection of the

null hypothesis and concludes that there is a significant

difference between the assessment of experts and users.

Looking back at the comparison of the assessments, the

experts’ assessment is greater than the users’

assessment. This means that the system is appreciated by

experts from a technical view but not much by the user.

The expert appreciates the performance efficiency of the

system compared to the user since the experts are most

knowledgeable by means of time behavior, resource

utilization, and capacity of the system. In addition, the

expert appreciates the usability of the system since the

69
experts learn features more quickly and retain their

knowledge longer, which directly correlates to decreased

training costs and time. Lastly, the expert appreciates

the reliability of the system compared to the user since

the system meets their expectation by means of it can

operate in a defined environment without failure.

70
 CHAPTER 5

Summary, Conclusion, and Recommendations

This chapter presents the summary of the research

work undertaken, the conclusions drawn, and the

recommendations made as an outgrowth of this study. The

findings presented in this report were based on the

investigation performed using specific materials. Some

limitations have been identified.

Summary of Findings

The salient findings of this study to the specific

question are as follows:

Problem 1. What is the Impact of Blockchain Technology to

Protect the Integrity of the Video and Image to prevent

Identity Theft using Deepfake Analyzer?

With emerging technology initially alarming since

image manipulation, and face-swapping create a

massive threat to violation of privacy and identity,

researchers were driven to develop a system that

will identify the impact of blockchain technology to

protect video and image integrity from identity

71
 theft using a deepfake analyzer. Table 3 displays

the data acquired in the pre-survey of the study

determining the Impact of blockchain technology to

protect video and image integrity from identity

theft using a deepfake analyzer. The series of

questions is strongly supported by all three groups

of responders. The pre-survey suggests that there is

a problem on Deepfake that must be addressed and

fixed.

Problem 2. How do the respondents evaluate “Deepfake

analyzer using Blockchain technology” in terms of the

following:

a. Functional Suitability

b. Performance Efficiency

c. Compatibility

d. Usability

e. Reliability

f. Security; and

g. Maintainability

According to the research findings, the system

"agreed" with both users’ and expert respondents'

assessments in terms of Functional Suitability,

72
Performance Efficiency, Compatibility, Usability,

Reliability, Security, Maintainability, implying that the

system solved the problems and integrated all of the

features required for the system to function. In

addition, table 8 displays that among the eight features

of the system, the functional appropriateness evaluation

had the highest weighted mean of 4.68 and was interpreted

as agree by the respondents. On the contrary, the

usability evaluation had the lowest weighted mean of 4.47

and was read as agree by the respondents.

Conclusion

Nowadays, we hear about current AI developments,

particularly deepfakes, and see terrifying news about the

use of deepfake technology against famous personalities

and politicians. We all know that invasion of privacy can

harm people and this behavior is not acceptable. Based on

the indicated findings the researchers concluded the

following statement of the problems.

1. What is the Impact of Blockchain Technology to

Protect the Integrity of the Video and Image to

prevent Identity Theft using Deepfake Analyzer?

73
 Based on the findings, the result shows that the

problem that drove the researcher in creating “deepfake

analyzer to protect the integrity of the video and image

to prevent identity theft” are “Do you agree that

Deepfake can be used to spread fake news and lead to an

increase in fake scandals”, “Do you agree that deepfake

can be misused to commit fraud, extortion, bullying as

well as to falsify evidence” and “it is important to

fight against false information and to take an action

about Deepfake”. These are the factors or the problems

that should be solved by the developed system.

2. How do the respondents evaluate “Deepfake analyzer

using Blockchain technology” in terms of the

following:

- Functional Suitability

- Performance Efficiency

- Compatibility

- Usability

- Reliability

- Security

- Maintainability

74
Based on the findings of the research, it showed that the

system was assessed as “agree” by the respondents both

users and experts in terms of Functional Suitability,

Performance Efficiency, Compatibility, Usability,

Reliability, Security, Maintainability, and can be

interpreted that the system successfully solved the

problems and integrated all of the feature needed for the

system to be functional.

Recommendation

Here are a few recommendations to further improve

and maximize the system's capabilities. The following

list should be regarded with caution, as no comprehensive

research has been conducted to determine whether or not

they are viable in this circumstance.

1. The study shows that the system can protect the

integrity of the Video and Image to prevent

Identity Theft using Deepfake Analyzer. The

researchers recommend adding audio detection to

be studied another feature for deepfake

detection.

75
2. Future researchers can focus on improving the

system user’s usability as this feature receive

the lowest evaluation by means of appropriateness

recognizability, learnability, operability, user

error protection, user interaction aesthetic, and

accessibility.

3. Future studies can broaden the system's use and

increase the size of the sampled data for the

system.

76
 REFERENCES:

Akwei, A. (2022). o g n a M o g n a M o g n a M o g n a
M. https://44bars.com/andrea-brillantes-viral-
deepfake-video-is-the-new-internet-obsession/

Bgen, C., Crame, R. T., & City, Q. (n.d.). PHILIPPINE

NATIONAL POLICE ANTI-CYBERCRIME GROUP.
https://www.mondaq.com/canada/copyright/687716/what-
can-the-law-do-

Chauhan, A., & Kumar, A. (n.d.). Establishing Environment

Setup for Preventing Deepfakes using Blockchain
Technology.

Cheikosman, E., Analyst, P., Impact, C., Forum, W. E.,

Hewett, N., Lead, P., Technology, D. L., Forum, W.
E., Gabriel, K., Services, H. D., & Electronica, A.
(2021). Blockchain can help combat the threat of
deepfakes. Here’s how.

Ge, S., Lin, F., Li, C., Zhang, D., Tan, J., Wang, W., &
Zeng, D. (2021, December 1). Latent Pattern Sensing:
Deepfake Video Detection via Predictive
Representation Learning. ACM International Conference
Proceeding Series.
https://doi.org/10.1145/3469877.3490586

77
Hasan, H. R., & Salah, K. (2019). Combating Deepfake
Videos Using Blockchain and Smart Contracts. IEEE
Access, 7, 41596–41606.
https://doi.org/10.1109/ACCESS.2019.2905689

Li, X., Lang, Y., Chen, Y., Mao, X., He, Y., Wang, S.,
Xue, H., & Lu, Q. (2020). Sharp Multiple Instance
Learning for DeepFake Video Detection. MM 2020 -
Proceedings of the 28th ACM International Conference
on Multimedia, 1864–1872.
https://doi.org/10.1145/3394171.3414034

Li, Y., & Lyu, S. (n.d.). Exposing DeepFake Videos By

Detecting Face Warping Artifacts.

Lugstein, F., Baier, S., Bachinger, G., & Uhl, A. (2021).

PRNU-based Deepfake Detection. IH and MMSec 2021 -
Proceedings of the 2021 ACM Workshop on Information
Hiding and Multimedia Security, 7–12.
https://doi.org/10.1145/3437880.3460400

Mamunur Rashid, M., Lee, S.-H., & Kwon, K.-R. (2021).

Blockchain Technology for Combating Deepfake and
Protect Video/Image Integrity. Journal of Korea
Multimedia Society, 24(8), 1044–1058.
https://doi.org/10.9717/kmms.2021.24.8.1044

Nagothu, D., Xu, R., Chen, Y., Blasch, E., & Aved, A.
(2022). DeFake: Decentralized ENF-Consensus Based
DeepFake Detection in Video Conferencing. 1–6.
https://doi.org/10.1109/mmsp53017.2021.9733503

News, S. (2022). What are Deepfakes and Fake News? Listen

78
 to this Story What are deepfakes?
https://sulit.ph/what-are-deepfakes-and-fake-news/?
fbclid=IwAR1oxogpwdZ4_tXCBX-PQdcrI-
Nkoqg5ru_NmkOfE0hbbYJ-VpCgy-2nfYE

Patil, U., & Chouragade, P. M. (2021a). Deepfake Video

Authentication Based on Blockchain. Proceedings of
the 2nd International Conference on Electronics and
Sustainable Communication Systems, ICESC 2021, 1110–
1113. https://doi.org/10.1109/ICESC51422.2021.9532725

Patil, U., & Chouragade, P. P. M. (2021b). Blockchain

Based Approach for tackling Deepfake videos.
International Journal of Scientific Research in
Computer Science, Engineering and Information
Technology, 342–347.
https://doi.org/10.32628/cseit217372

Tambe, S. N., & Pawar, A. B. (2020). Detecting fake

Videos using Block Chain and Smart Contracts.
International Journal of Recent Technology and
Engineering, 8(4S5), 209–213.
https://doi.org/10.35940/ijrte.d1049.1284s519

79
 APPENDIX A

Plagiarism Test Result

80
 APPENDIX B

Communication Letters

81
 APPENDIX C

Sample Collected Forms and Reports Pre-Survey

82
Figure. 12 Age of the Respondents

Figure. 13 Question No. 1

Figure. 14 Question No. 2

Figure. 15 Question No. 3

Figure. 16 Question No. 4

83
Figure. 17 Question No. 5

84
Figure. 18 Question No. 6

Figure. 19 Question No. 7

85
Figure. 20 Question No. 8

Figure. 21 Question No. 9

86
Figure. 22 Question No. 10

87
 APPENDIX D

Sample Collected Forms and Reports Post-Survey

Figure. 23 Age of Respondents

Figure. 24 Birth Year of Respondents

88
Figure. 25 Functional Suitability in terms of Functional

Completeness

Figure. 26 Functional Suitability in terms of Functional

Correctness

89
Figure. 27 Functional Suitability in terms of Functional

Appropriateness

Figure. 28 Performance Efficiency in terms of Time

Behavior

90
Figure. 29 Performance Efficiency in terms of Resource

Utilization

Figure. 30 Performance Efficiency in terms of Capacity

91
 Figure. 31 Compatibility in terms of Co-existence

Figure. 32 Compatibility in terms of Interoperability

92
Figure. 33 Usability in terms of Appropriateness

Recognizability

Figure. 34 Usability in terms of Learnability

93
 Figure. 35 Usability in terms of Operability

Figure. 36 Usability in terms of User Error Protection

94
Figure. 37 Usability in terms of User Interface

Aesthetics

Figure 38. Usability in terms of Accessibility

95
 Figure. 39 Reliability in terms of Maturity

Figure. 40 Reliability in terms of Availability

96
Figure. 41 Reliability in terms of Recoverability

Figure. 42 Security in terms of Confidentiality

97
 Figure. 43 Security in terms of Integrity

Figure. 44 Security in terms of Non-Repudiation

98
Figure. 45 Maintainability in terms of Modifiability

Figure. 46 Maintainability in terms of Testability

99
 APPENDIX E

Figure. 47 User Interface for Image Upload and

Prediction

100
Figure. 48 User Interface for Video Upload and

Prediction

Figure. 49 User Interface for the Result

101
 APPENDIX F

User Manual

Figure. 50 Select a photo that you want to test

102
 Figure. 51 Generation of Hash Code

Figure. 52 Testing of Altered Image to see the Accuracy

Figure. 53 Result of the Test

103
 Figure. 54 Testing of Video

Figure. 55 Generation of Hash Code

104
Figure. 56 History of the Previous Result

Figure. 57 Delete Button

105
 APPENDIX G

Curriculum Vitae

Personal Information

Mathew Ern V. Aquilizan

1 Lilac Street Area 9B Veterans

Village Q.C

09777629694

[email protected]

Academic History
Eulogio “Amang” Rodriguez Institute of (2019 – Present)
Science and Technology
Scholar: First year- Present

Senior High School: New Era University (2017 – 2019)

Awards and Honors

Student Assistant Office of the President (2022-2023)

Chairman Commission on Student Election (2022-2023)
President Computer Science Association (2021-2022)
CAS Commissioner Commission on (2021-2022)
Student Election
President Youth Organization (2021-2022)
Class Mayor (2019 – Present)
CAS Athlete (2019-2020)

106
Senior High Graduate with Highest Honor (2018-2019)

Personal Information

John Angelo C. Manuel

09167940894

2 G. Fernando St. Malanday Marikina City

[email protected]

A young aspiring developer who aims to learn more.

Academic History
Eulogio “Amang” Rodriguez Institute of (2019 – Present)
Science and Technology
Scholar: First year- Present

SHS: Christian Life College (2017 – 2019)

Awards and Honors

Class President (2017- 2018)

SSG Officer (2017- 2019)

Journalist (2017- 2019)

NCLC News Team Editor (2017- 2019)

Class Vice President (2018- 2019)

107
Personal Information

Kyle Vincent L. Merencilla

148 -A Osmeña St. Tondo Manila.

09998099695

[email protected]

Academic History

Eulogio “Amang” Rodriguez Institute of (2019 – Present)

Science and Technology
Scholar: First year- Present

SHS: Tondo High School Manila, (2017 – 2019)

Awards and Honors

SSG Officer (2015 - 2016)

Class Vice President (2016 - 2017)

Class Vice President (2018 – 2019)

Class Vice Mayor (2020 – 2021)

Class Vice Mayor (2021 – 2022)

108
Personal Information

Liezl Leine P. Perez

750 Sunstone Street, Palmera 2,
Taytay, Rizal
09772403961
perezliezl.bscs@gmail. com

A hard- working, motivated, and optimistic student who

takes up responsibilities with utmost enthusiasm.
Academic History
Eulogio “Amang” Rodriguez Institute of (2019 – Present)
Science and Technology
Scholar: First year- Present

SHS: Juan Sumulong Memorial Junior Colleg(2017 – 2019)

Awards and Honors

Secretary: Supreme Student Council (2016- 2017)

With Honors: Grade 11 (2017- 2018)

President: Teatro Sumulong (2018- 2019)

Member: JSMJC SHS Chorale (2018- 2019)

Art Representative: Supreme Student Council (2018- 2019)

With Honors: Grade 12 (2018- 2019)

Player: EARIST Taekwondo (2019)

109
Personal Information
Raffy D. Uao
Address: Block 28 Unit 6 Sitio Sto. Niño
Mayamot Antipolo City
09569285131
[email protected]

Educational Background
Eulogio Amang Rodriguez Institute of (2019–Present)
Science and Technology
SHS: Ama East Rizal (2016-2018)
High School: Antipolo National High School (2012-2016)
Elementary: Bagong Nayon 1 Elementary School (2007 –
2012)
Elementary: Cupang Elementary School (2006–2007)

110
111