DEEPFAKE DETECTION

An Industry Oriented Mini Project Report Submitted to

Jawaharlal Nehru Technological University Hyderabad

In partial fulfillment of the

requirements for the award of the
degree of
BACHELOR OF TECHNOLOGY
IN
COMPUTER SCIENCE AND ENGINEERING

By
A.ROHIT 22E11A6201
D.SAI SHIVA RAM REDDY 22E11A6213
K.ESHWAR 22E11A6228
N.VENKATESH 22E11A6241
P.MANIDEEP 22E11A6245

Under the guidance of

DR.Ms.Nazneen Fathima
Assistant Professor
Department of Computer Science and Engineering

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

BHARAT INSTITUTE OF ENGINEERING AND TECHNOLOGY
Accredited by NAAC, Accredited by NBA (UG Programmes: CSE, ECE, EEE & Mechanical)
Approved by AICTE, Affiliated to JNTUH Hyderabad
Ibrahimpatnam -501 510, Hyderabad, Telangana
JUNE 2025
 DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
BHARAT INSTITUTE OF ENGINEERING AND TECHNOLOGY
Accredited by NAAC, Accredited by NBA (UG Programmes: CSE, ECE, EEE & Mechanical)
Approved by AICTE, Affiliated to JNTUH Hyderabad
Ibrahimpatnam -501 510, Hyderabad, Telangana

Certificate
This is to certify that the Industry Oriented Mini Project work
entitled
“DEEPFAKE DETECTION” is the bonafide work done
By

A.ROHIT 22E11A6201
D.SAI SHIVA RAM REDDY 22E11A6213
K.ESHWAR 22E11A6228
N.VENKATESH 22E11A6241
P.MANIDEEP 22E11A6245
in the Department of Computer Science and Engineering, BHARAT
INSTITUTE OF ENGINEERING AND TECHNOLOGY, Ibrahimpatnam is
submitted to Jawaharlal Nehru Technological University,
Hyderabad in partial fulfillment of the requirements for the award of
B.Tech degree in Computer Science and Engineering during 2024-
2025.

Supervisor: Dr.Ms.Nazneen Fathima Department I/C Dr.Ms.Nazneen Fathima

Assistant Professor Assistant Professor
Dept of Computer Science and Engineering Dept of Computer Science and Engineering
Bharat Institute of Engineering and Technology Bharat Institute of Engineering and Technology
Ibrahimpatnam–501 510, Hyderabad Ibrahimpatnam– 501 510, Hyderabad
Viva-Voce held on……………………………………………

Internal Examiner External Examiner

ACKNOWLEDGEMENT

The satisfaction that accompanies the successful completion of the task would be put
incomplete without the mention of the people who made it possible, whose constant guidance
and encouragement crown all the efforts with success.

We avail this opportunity to express our deep sense of gratitude and hearty thanks to Sri
CH. Venugopal Reddy, Chairman & Secretary of BIET, for providing congenial atmosphere
and encouragement.

We would like to thank Prof. G. Kumaraswamy Rao, Former Director & O.S. of DLRL
Ministry of Defence, Sr. Director R&D, BIET, and Dr. V Srinivasa Rao, Dean CSE, for having
provided all the facilities and support.

We would like to thank our Department Incharge / HOD Ms.Nazneen Fathima, for
encouragement at various levels of our Project.

We are thankful to our Project Coordinator Ms.Nazneen Fathima, Assistant Professor,

Computer Science and Engineering for her support and cooperation throughout the process of
this project.

We are thankful to our guide Ms.Nazneen Fathima, Assistant Professor, Computer

Science and Engineering for his sustained inspiring Guidance and cooperation throughout the
process of this project. His wise counsel and suggestions were invaluable.

We express our deep sense of gratitude and thanks to all the Teaching and Non-Teaching
Staff of our college who stood with us during the project and helped us to make it a successful
venture.

We place highest regards to our Parent, our Friends and Well-wishers who helped a lot in
making the report of this project
A.ROHIT [22E11A6201]
D.SAI SHIVA RAM REDDY [22E11A6213]
K.ESHWAR [22E11A6228]
N.VENKATESH [22E11A6241]
P.MANIDEEP [22E11A6245]
 DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
BHARAT INSTITUTE OF ENGINEERING AND TECHNOLOGY
Accredited by NAAC, Accredited by NBA (UG Programmes: CSE, ECE, EEE & Mechanical)
Approved by AICTE, Affiliated to JNTUH Hyderabad
Ibrahimpatnam -501 510, Hyderabad, Telangana

Declaration
We hereby declare that this Industry Oriented Mini Project is titled DEEPFAKE

DETECTION is a genuine Industry Oriented Mini Project work carried out by us, in
B.Tech {Computer Science and Engineering (CYBER SECURITY)}
degree course of Jawaharlal Nehru Technology University
Hyderabad, Hyderabad and has not been submitted to any other course or
university for the award of my degree by me.

Signatures of the Project team members

1.

2.

3.

4.

5.
 ABSTRACT

This Industry Oriented Mini Project titled “DEEP FAKE” is a web application focuses on
identifying manipulated human faces in images or videos created using advanced deep
learning techniques like GANs and autoencoders. These forgeries often exhibit subtle
inconsistencies in facial features, expressions, eye blinking, head movements, or skin texture.
Detection methods leverage convolutional neural networks (CNNs) to analyze spatial and
temporal anomalies in facial regions. Techniques such as frequency analysis and
physiological cue detection further enhance accuracy. Datasets like FaceForensics++ and
Celeb-DF are commonly used for training. Effective facial deepfake detection is critical for
preventing misinformation, identity fraud, and digital impersonation.

Keywords: Deepfake Detection, Facial Manipulation

{
P
 TABLE OF CONTENTS

Chapter Page
Title
No. No.
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Abstract.............................................................................................................................................................................. v

Table of Contents..............................................................................................................................................................vi

List of Figures..................................................................................................................................................................viii

List of Tables......................................................................................................................................................................ix

List of Symbols and Abbreviations.....................................................................................................................................x

1 Introduction………………………………………………………………………………………… 1
1.1 Android Technical Environment……………………………………………… 1

1.2 App Components………………………………………………..…………………… 2

1.3 Activating components…………………………………………….……………… 4 1.4 Introduction to deepfake

detection…………………………………………………………..

2 Related Work……………………………………………………………………………………….. 7
3 Motivation……………………………………………………………………………………………. 8
4 Objectives……………………………………………………………………………………………. 9
5 Problem Statement……………………………………………………………………………… 11
6 Design Methodology……………………………………………………………………………… 12

6.1 Architecture……………………………………………………………………………………. 12

6.2 Modules………………………………………………………………………………………… 14

6.3 Requirements Specifications…………………………………………………………. 21

6.3. UML Diagrams………………………………………………………………………………. 22

7 Experimental Studies…………………………………………………………………………… 28

P
7.1 Test Cases……………………………………………………………………………………… 31

7.2 Results Analysis…………………………….………………………………………………. 32

{
8 Conclusion and Future Scope………………………………………………................................ 36

References………………………………………………………………………………………………………. 37

{
P
 LIST OF FIGURES

Figure Caption Page No.

No.
6.1 System Analysis Architecture……………………………………………………………….. 12

6.2 Technical architecture of proposed system ………………………………………………. 13

6.3 Use case diagram for deepfake detection……………………………………………………………. 22

6.4 Class diagram for deepfake detection ………………………………………………………………….. 23

{
PA
{
PA
LIST OF SYMBOLS AND ABBREVIATIONS

LIST OF ABBREVIATIONS
Symbol Description
APK Android Application Package
RFID Radio Frequency identification
OMA Open Mobile Alliance
DS Data Synchronization

{
PA
 2024-2025

1. INTRODUCTION
1.1. Overview of Deepfakes
Deepfakes are synthetic media, such as images or videos, created using artificial
intelligence (AI) and machine learning (ML) techniques. These manipulated media can
be incredibly realistic, making it challenging to distinguish between authentic and fake
content.
1.2.Impact of Deepfakes
Deepfakes have significant implications for various aspects of society, including:
- Misinformation and disinformation: Deepfakes can be used to spread false information,
potentially influencing public opinion or causing harm to individuals or organizations.
- Identity theft and impersonation: Deepfakes can be used to impersonate individuals,
potentially leading to identity theft or reputational damage.
1.3.Importance of Detection
Detecting deepfakes is crucial to mitigate their potential harm. Effective detection methods can
help:

- Prevent the spread of misinformation

- Protect individuals' identities and reputations
- Maintain trust in digital media

2. RELATED WORK
Existing Deepfake Detection Methods
Several methods have been proposed for detecting deepfakes, including:

- Machine learning-based approaches: These methods use ML algorithms to classify media as

real or fake.
- Deep learning-based approaches: These methods use deep neural networks to detect
deepfakes.

Facial Detection in Deepfake Detection

{ PAGE } |
Page
 2024-2025
Facial detection plays a crucial role in deepfake detection, as it can help identify manipulated
faces in images or videos.

Limitations of Current Methods

Current deepfake detection methods have limitations, including:

- Limited accuracy: Detection methods may not always accurately distinguish between real
and fake media.
- Vulnerability to adversarial attacks: Detection methods can be vulnerable to adversarial
attacks, which can compromise their effectiveness..

3. MOTIVATION
Need for Accurate Detection
There is a pressing need for accurate and reliable deepfake detection methods to
mitigate the potential harm caused by deepfakes.

Potential Applications
Deepfake detection has various potential applications, including:

- Cybersecurity: Deepfake detection can help prevent cyber attacks that utilize
deepfakes.
- Law enforcement: Deepfake detection can aid law enforcement agencies in
investigating crimes involving deepfakes.

4. OBJECTIVES
Primary Objective

The primary objective of this project is to detect deepfakes using facial detection.

Secondary Objectives

The secondary objectives include:

- Improving detection accuracy

- Enhancing robustness and efficiency

{ PAGE } |
Page
 2024-2025
4.1 FEASIBILITY STUDY:
The feasibility of the project is analyzed in this phase. “All projects are feasible given
unlimited resources and infinite time”. But in reality both resources and time are scarce.
Project should confirm to time bounce and should be optimal in there consumption of
resources. These places a constant are approval of any project. Feasibility has applied to
Maintenance of Elementary School Data pertains to the following areas:

• TECHNICAL FEASIBILITY

• OPERATIONAL FEASIBILITY

• ECONOMICAL FEASIBILITY

TECHNICAL FEASIBILITY:

To determine whether the proposed system is technically feasible, we should take into
consideration the technical issues involved behind the system. Maintenance of Elementary
School Data uses the web technologies, which is rampantly employed these days worldwide.
The world without the web is incomprehensible today. That goes to proposed system is
technically feasible.

OPERATIONAL FEASIBILITY:

To determine the operational feasibility of the system we should take into consideration the
awareness level of the users. This system is operational feasible since the users are familiar
with the technologies and hence there is no need to gear up the personnel to use system.
Also the system is very friendly and to use.

ECONOMIC FEASIBILITY

To decide whether a project is economically feasible, we have to consider various factors

as:

• Cost benefit analysis

• Long-term returns

{ PAGE } | P
age
 2024-2025
• Maintenance costs

The proposed Maintenance of Elementary School Data is computer based. It requires

average computing capabilities and access to internet, which are very basic requirements
and can be afforded by any organization hence it doesn’t incur additional economic
overheads, which renders the system economically feasible.

5. PROBLEM STATEMENT

Deepfake Detection Challenge

Detecting deepfakes is a challenging task due to the sophistication of deepfake generation
methods.

Facial Detection in Deepfakes

Using facial detection for deepfake detection poses specific challenges, including:

- Identifying manipulated faces

- Distinguishing between real and fake facial expressions

5.1 Existing System:

In the existing system, now a days finding the right application has become a tedious task and
to use them with ease of access is not the status quo. Lot of precious productive time is
perhaps wasted upon the pursuit of finding a free app which facilitates our needs with
efficiency and ease-of-access. Existing system doesn’t have display of the medicine by using
images along with the alarm system. Users have to pay money to use the facilities provided by
the application and it needs a lot of hardware equipments when implemented using
IOT(internet of things).

5.2 Proposed System:

System Components

- Facial Detection: The system uses facial detection algorithms, such as MTCNN Face
detector, to identify faces in images or videos.

- Deepfake Detection Model: The detected faces are then fed into a deepfake detection model,
which classifies them as real or fake. This model can be based on various architectures, such
as convolutional neural networks (CNNs) or vision transformers (ViTs).

{ PAGE } |
Page
 2024-2025
- Patch-Wise Deep Learning Model: Another approach uses a patch-wise deep learning model,
which analyzes specific regions of the face to detect deepfakes ¹ ².

Techniques Used

- Single-Modal Detection: This involves analyzing a single modality, such as images or

videos, to detect deepfakes.

- Multi-Modal Detection: This approach combines multiple modalities, such as audio and
visual information, to improve detection accuracy.

- Frequency Analysis: Some systems use frequency analysis techniques to identify forgery
patterns in images or videos.

- Dynamic Face Augmentation: This technique involves augmenting face data to improve the
robustness of deepfake detection models

6. DESIGN METHODOLOGY

6.1 SYSTEM ARCHITECTURE

The System architecture consists of all the modules of the main module (Utility Shack.exe)
are shown below. However, it also shows how the modules access the resources and how
they interact with each other and communicate with each other. It also shows the input,
output and processing pathways of the application and also the flow of direction of input,
output and processing directions of the application.

{ PAGE } | P
age
 2024-2025

Figure 6.1: System analysis architecture

6.2 MODULES:
6.2.1 Home page:
Home page is the First page visible for the user which contains two buttons named:
i. REGISTER BUTTON
ii. LOGIN BUTTON
If the user have no account in the application then the user has to press the “Register
Button” and if the user already have an account he then press the button named “Login”.

Abstract Source code:

from flask import Flask,
render_template, request
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.image img_to_array,
load_img
import numpy as np
import os
app = Flask(name)
model = load_model('deepfake_detection_model.h5’)

{ PAGE } |
Page
 2024-2025
@app.route('/’)
def index():
return render_template('index.html’)
@app.route('/predict', methods=['POST’])
def predict():
image = request.files['image']
image_path = os.path.join('static', image.filename)
image.save(image_path)
img = load_img(image_path, target_size=(256, 256))
img = img_to_array(img) / 255.0
img = np.expand_dims(img, axis=0)
prediction = model.predict(img)[0][0]
result = "Fake" if prediction > 0.5 else "Real“
return render_template('index.html', result=result, image=image_path)
if name == 'main':
app.run(debug=True)

6.2.3 LOGIN PAGE:

If the user already have an account in the application then the user has to click the login
button in the home page and the application navigates the user to the “Login page”.

6.2.4 Admin Home Page:

Admin is the person who have control over all the users in the application and also have
control on the database.

6.3 Requirement Specifications:

6.3.1 Minimum Software Requirements

 Operating systems : WINDOWS 7/10/XP

 IDE ; ANDROID STUDIO SDK VERSION 3
 Back end : JAVA SE 11.0.3 jdk-8u231-windows-x64
 Front end : XML VERSION 1.0

 Database : FIREBASE 6.11.0

{ PAGE } | P
age
 2024-2025
6.3.2 Minimum Hardware Requirements
• Processor : Intel i9
• RAM Capacity : 512 MB (or) Higher
• Cache : 256 MB
• Hard Disk : 10 GB

{ PAGE } |
Page
 2024-2025

6.4 UML diagrams:

6.4.1 Use case diagram for deepfake detection

Figure 6.3: Use case diagram for deepfake detection

A use case diagram for medping shows as set of use cases and actors (admin ,user) and their
relationships .This Use case diagram is especially important in organizing and modeling
behavior of a medping system. It shows the action that are to be performed by the both the
user and admin.

{ PAGE } | P
age
 2024-2025
6.4.2 Class diagram for deepfake detection

Figure 6.4: Class diagram for deepfake detection

This class diagram shows the set of classes, interfaces, collaboration and their relationships.
This Class diagram for medping includes all the active classes that are used to address this
static process view of a system.

{ PAGE } |
Page
 2024-2025
7. EXPERIMENTAL STUDIES

7.1 Testing Process:

Black-box Testing and White-box Testing:

Black box testing is the Software testing method which is used to test the software without
knowing the internal structure of code or program. White box testing is the software
testing method in which internal Structure is being known o tester who is going to test the
software.

A strategy for system testing integrates system test cases and design techniques into a well
planned series of steps that results in the successful construction of software. The testing
strategy must cooperate test planning, test case design, test execution, and the resultant
data collection and evaluation. A strategy for software testing must accommodate lowlevel
tests that are necessary to verify that a small source code segment has been correctly
implemented as well as high level tests that validate major system functions against user
requirements.

Software testing is a critical element of software quality assurance and represents the
ultimate review of specification design and coding. Testing represents an interesting
anomaly for the software. Thus, a series of testing are performed for the proposed system
before the system is ready for user acceptance testing.

Testing Methodologies:

The following are the testing Methodologies:

• Unit Testing

• Integration Testing

• User Acceptance Testing

• Output Testing

• Validation Testing

{ PAGE } | P
age
 2024-2025

Unit Testing:

Unit testing is essentially for the verification of the code produced during the coding phase
and the goal is test the internal logic of the module/program. In the Generic code project,
the unit testing is done during coding phase of data entry forms whether the functions are
working properly or not. In this phase all the drivers are tested they are rightly connected or
not.

Integration Testing:

All the tested modules are combined into sub systems, which are then tested. The goal is to
see if the modules are properly integrated, and the emphasis being on the testing interfaces
between the modules. In the generic code integration testing is done mainly on table
creation module and insertion module.

The following are the types of Integration Testing:

1. Top Down Integration

2. Bottom up Integration

User Acceptance Testing:

User acceptance of a system is the key factor for the success of any system. The system
under consideration is tested for user acceptance by constantly keeping in touch with the
prospective system users at the time of developing and making changes whenever required.
The system developed provides a friendly user interface that can easily be understood even
by a person who is new to the system.

Output Testing:

After performing the validation testing the next step is output testing of the proposed
system, since no system could be useful if it does not produce the required output in the
specified format.
Asking the users about the format required by them tests the outputs generated or displayed
by the system under consideration. Hence output format is considered in 2 ways- one is on
screen and another in printed format.

{ PAGE } |
Page
 2024-2025

Validation Testing:

This testing concentrates on confirming that the software is error- free in all respects. All
the specified validations are verified and the software is subjected to hard-core testing. It
also aims at determining the degree of deviation that exists in the software designed from
the specification; they are listed out and are corrected.

8. CONCLUSION AND FUTURE WORK

CONCLUSION:
In conclusion, the proposed system for deepfake detection using facial detection has shown
promising results in identifying manipulated media. By leveraging facial detection algorithms
and deep learning models, this system can effectively distinguish between real and fake
content.

Key Findings
The study has demonstrated the effectiveness of using facial detection for deepfake detection,
with the proposed system achieving high accuracy in identifying manipulated faces. The use
of patch-wise deep learning models and frequency analysis techniques has also shown
potential in improving detection accuracy.

Implications
The findings of this study have significant implications for various fields, including
cybersecurity, law enforcement, and social media. By developing effective deepfake detection
methods, we can mitigate the potential harm caused by manipulated media and maintain trust
in digital content.

{ PAGE } | P
age
 2024-2025

FUTURE WORK

1. Improving Robustness: Developing more robust deepfake detection methods that can
withstand adversarial attacks and various types of manipulation.

2. Real-Time Detection: Creating real-time deepfake detection systems that can identify
manipulated media in real-time.

3. Multi-Modal Detection: Exploring multi-modal detection approaches that combine

visual, audio, and other types of data to improve detection accuracy.

4. Explainability and Transparency: Developing methods to provide explanations and

insights into deepfake detection decisions, improving transparency and trust.

5. Large-Scale Evaluation: Conducting large-scale evaluations of deepfake detection

methods on diverse datasets to assess their performance and generalizability.

6. Adversarial Attack Detection: Developing methods to detect and mitigate adversarial

attacks designed to evade deepfake detection.

7. Integration with Existing Systems: Integrating deepfake detection with existing

cybersecurity and social media systems to prevent the spread of manipulated media.

Potential Applications

1. Social Media Monitoring: Using deepfake detection to monitor and mitigate the spread
of manipulated media on social media platforms.

2. Cybersecurity: Integrating deepfake detection with cybersecurity systems to prevent

cyber attacks that utilize deepfakes.

3. Law Enforcement: Assisting law enforcement agencies in investigating crimes involving

deepfakes.

4. Content Authentication: Developing methods to authenticate the authenticity of digital

content, ensuring its integrity and trustworthiness.

By exploring these future work directions, researchers and developers can continue to
improve the effectiveness and robustness of deepfake detection methods.

{ PAGE } |
Page