FINAL YEAR PROJECT

PROPOSAL ON

IMAGE ENCRYPTION AND DECRYPTION USING ADVANCED CRYPTOGRAPHIC

ALGORITHM

PRESENTED BY

OMONIYI WARIZ OLOLADE

CSC/2020/1149

SUBMITTED TO SUPERVISOR:

DR. AWOYEMI

IN PARTIAL FUFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF

BACHELOR OF COMPUTER SCIENCE (B.Sc COMPUTER SCIENCE)

DATE: NOVEMBER, 2024

 INTRODUCTION
The security and privacy of sensitive information, including images, have become paramount
due to the widespread integration of technology in various aspects of our lives. As individuals,
organizations, and governments exchange vast amounts of visual data online, concerns regarding
the unauthorized access, tampering, and interception of these images have escalated
significantly. This has led to a heightened emphasis on employing robust security measures to
protect the confidentiality and integrity of such data (Chanal et al., 2020). Image encryption and
decryption, both integral components of modern cryptography, play a pivotal role in addressing
these concerns.
Cryptography is the science of secure communication, encompassing techniques to convert
information into unreadable formats, which can only be converted back to a readable form by
authorized individuals possessing the appropriate decryption keys. Image encryption deals
explicitly with the transformation of visual data, such as photographs, diagrams, and medical
scans, into encrypted formats that are challenging to decipher without the correct keys.
Encryption ensures that even if an unauthorized party gains access to the encrypted images, they
would be unable to comprehend the content without the decryption key.
This level of protection extends not only to stored images but also to images transmitted over
digital networks, making it significantly harder for attackers to intercept and understand the
information being exchanged (Alkhalil et al. 2021). Consider the case of medical imaging.
Hospitals frequently transmit sensitive medical images, such as X-rays and MRIs, between
departments or to specialists for diagnosis. These images contain private patient information and
must be kept confidential. Image encryption ensures that only authorized medical professionals
with the appropriate decryption keys can access and interpret these images, preventing
unauthorized personnel from obtaining sensitive patient data.
In industries like e-commerce and finance, image encryption is vital for securing financial
transactions and customer data. Websites that deal with user-generated content, such as social
media platforms, also rely on encryption to protect users' images from malicious actors. One
prominent encryption technique used in image security is the application of cryptographic
algorithms like AES (Advanced Encryption Standard) to pixel data. These algorithms are
mathematically designed to be computationally secure, making it extremely difficult for
unauthorized individuals to reverse-engineer the encryption process and obtain the original
image.

This work will not only secure image data but will also set the foundation for scalable solutions
that can be integrated into real-world applications, such as secure image sharing platforms and
confidential communication systems.

AIM AND OBJECTIVES

AIM

The project aims to analyze and enhance image encryption and decryption techniques,
investigate vulnerabilities in existing algorithms, develop efficient solutions, and explore
innovative methodologies to ensure secure visual data transmission and storage in today's
dynamic digital environment. This system should meet the following key criteria:

i. Robust Security: Ensure that encrypted images are highly resistant to unauthorized
access and various forms of cryptanalysis.
ii. Efficiency: The system should process encryption and decryption quickly, even for
large and high-resolution images, without significant degradation in performance.
iii. Scalability: The system should be scalable, capable of handling diverse types of
images, from small grayscale images to complex, high-resolution color images.
iv. Practical Usability: The solution should be user-friendly and integrable with existing
digital systems, such as secure file-sharing platforms or online storage services.

OBJECTIVES
The objectives of this study are to:

i. Designing an Advanced Cryptographic Framework:

 Develop a robust encryption framework tailored to the characteristics of

image data.
  Combine symmetric encryption (for speed) and asymmetric encryption
(for security).

ii. 2. Implementing the Cryptographic System:

 Implement encryption algorithms (AES), optimized for image data.

 Use software platforms, Python for implementation.

3. Evaluating System Performance:

 Test the system using performance metrics

 Compare the system against existing solutions to evaluate improvements
in speed, security, and usability.

4. Optimizing for Scalability:

 Ensure the system can process various image formats (JPEG, PNG, BMP).

STATEMENT OF PROBLEM
In the modern digital landscape, safeguarding image data has become crucial due to the frequent
exchange of images across various platforms. While traditional encryption techniques work well
for text, they often fall short when dealing with the complex nature and larger size of image files.
The growing threats of cyberattacks, particularly in fields like healthcare and multimedia,
highlight the need for more sophisticated methods to secure visual data.

This research explores innovative strategies for image encryption and decryption, aiming to
address the limitations of current techniques. By tackling the unique challenges of securing
pixel-based image data, the study seeks to contribute to the development of more secure and
efficient encryption methods.

Key Issues:
1. Computational Overhead: Standard cryptographic algorithms, when applied to
large images, result in significant processing delays and resource consumption,
making them impractical for real-time applications.
 2. Vulnerability to Attacks: Simple encryption schemes are often susceptible to
brute-force attacks, differential cryptanalysis, and statistical analysis due to the
patterns present in image data.
3. Usability Challenges: Existing systems often lack user-friendly interfaces and
integration capabilities, limiting their adoption in everyday use cases, such as
secure image sharing.
4. Scalability Concerns: Many encryption algorithms struggle with scalability
when handling diverse image types, resolutions, and formats.

OVERVIEW OF CURRENT SYSTEMS

Current encryption methods for securing images fall into two primary categories: traditional
cryptographic algorithms and specialized image encryption techniques.

Traditional Cryptographic Algorithms:

1. AES (Advanced Encryption Standard): Widely used for data encryption, AES
offers strong security. However, it is computationally intensive for large image
data, leading to processing delays.
2. RSA (Rivest-Shamir-Adleman): RSA provides robust security through
asymmetric encryption but is inefficient for encrypting large image files due to its
reliance on computationally expensive operations.
3. ECC (Elliptic Curve Cryptography): ECC is a lightweight alternative to RSA,
offering similar security with lower resource consumption. However, it is less
commonly applied to image encryption.

Specialized Image Encryption Techniques:

1. Pixel Shuffling: Rearranges pixel positions in an image to obscure its content.

While fast, it is not secure on its own and must be combined with other
techniques.
 2. Chaotic Systems: Uses chaos theory to generate random encryption keys. While
secure, these systems can be difficult to implement and optimize.

OVERVIEW OF THE PROPOSED SYSTEM

The proposed system aims to overcome the challenges of existing methods by integrating
cryptographic algorithms with image-specific optimizations and steganographic techniques. It is
designed to ensure high security, efficiency, and adaptability to various use cases.

Key Features of the Proposed System:

1. Hybrid Encryption Approach:
• Combines symmetric encryption for fast processing with asymmetric
encryption to securely exchange encryption keys.
• Introduces dynamic key generation to enhance resistance to cryptographic
attacks.
2. Image-Optimized Encryption:
• Processes the unique structure of images (e.g., pixel correlation) to
minimize computational overhead.
• Preserves image quality during decryption, ensuring minimal loss of
fidelity.
3. Steganographic Integration:
• Allows for hidden transmission of sensitive images across unsecured
communication channels.
4. Scalability and Flexibility:
• Compatible with various image formats (JPEG, PNG) and resolutions.
• Supports integration with cloud platforms and mobile devices for real-time
image sharing and storage.
Workflow:

1. Encryption Process:
 • Input image is pre-processed (compressed if needed) to reduce size and
redundancy.
• Optionally, the encrypted image is embedded into a cover image using
steganography.
2. Decryption Process:
• The encrypted image is received by the intended recipient.
• Decryption keys are securely retrieved using asymmetric encryption
protocols.
• The original image is restored with minimal degradation.

IMPLEMENTATION STEPS
The implementation of the proposed system will follow a structured approach to ensure the
desired objectives are met.
Step 1: Requirement Analysis
• Identify the specific cryptographic needs for securing images, including
the target image formats, sizes, and resolutions.
• Determine the performance requirements (e.g., encryption speed, system
scalability).
Step 2: Algorithm Design
• Develop a hybrid cryptographic framework using AES for encryption and
ECC for secure key exchange.
• Design a steganographic module to embed encrypted images into cover
images.
• Implement adaptive algorithms to optimize processing for different image
types and resolutions.
Step 3: System Development
• Use programming languages Python for development.
• Create a user-friendly graphical interface for easy interaction with the
system.
Step 4: Testing and Evaluation
 • Functionality Testing: Verify that the system correctly encrypts and
decrypts images without errors.
• Performance Testing: Measure encryption and decryption time, especially
for high-resolution images.
• Security Testing: Assess resistance to attacks such as brute force,
differential cryptanalysis, and statistical attacks.
Step 5: Optimization
• Fine-tune algorithms to balance speed and security.
• Implement parallel processing techniques to improve performance for
large datasets.
• Ensure compatibility with low-resource environments.
Step 6: Deployment and Integration
• Test the system in real-world scenarios, such as secure image sharing
platforms.
• Integrate with cloud services for secure storage and retrieval of encrypted
images.

CONTRIBUTIONS TO KNOWLEDGE
This research contributes significantly to the field of image encryption by introducing a novel
key generation technique based on Gaussian distribution. This innovative approach offers a new
perspective on encryption methods, potentially enhancing security measures for image
protection. Additionally, the development of a practical desktop application using Python and
Tkinter provides a functional tool for secure image handling.
The study also introduces the use of Mean Squared Error (MSE) as a performance metric,
offering a quantitative measure of the system’s effectiveness in preserving image quality. These
contributions advance the understanding and application of image encryption technologies.

1. Novel Hybrid Cryptographic Framework:

 The project will introduce a hybrid encryption system that combines the speed of
symmetric algorithms (AES) with the security of asymmetric algorithms (RSA/ECC).
  This framework will be optimized specifically for image data, offering a tailored
solution for visual content security.
2. Integration of Steganography:
 This integration ensures that sensitive images can be transmitted discreetly, even over
unsecured channels.
3. Improved Performance Metrics:
 The system will demonstrate improved encryption and decryption times, making it
suitable for real-time applications.
 High-quality image restoration post-decryption will be achieved, ensuring minimal loss
of fidelity (high PSNR values).
4. Scalable and Practical Solution:
 The project will deliver a solution that can handle various image formats, resolutions, and
sizes.
 Compatibility with mobile devices will enhance its usability and scalability.
5. Advancing Research in Cryptography and Image Security:
 The findings and methodologies from this project will serve as a foundation for future
research in image security.
 The project will contribute to the development of encryption systems for other
multimedia data, such as videos and 3D models.
6. Real-World Applications:
 The system can be integrated into secure communication platforms, online storage
systems, and confidential document-sharing tools.
 It will benefit industries such as healthcare, defense, and social media by providing
robust image security solutions.

LITERATURE REVIEW

S/ Title Authors & Abstract Methodology Weakness/Limitation

N Year
i. Image encryption Kolivand, H, This paper The study uses The study identifies
techniques: A Hamood, SF, presents an various several limitations. One
comprehensive exhaustive methodologies major issue is the trade-
Asadianfam, S
review. Multimedia review of for encrypting off between
Tools and and Rahim,
research within images, computational efficiency
Applications
MS (2024) the field of focusing on and encryption strength,
image chaos-based with robust methods
encryption approaches. often being resource-
techniques. It The research intensive. Certain
commences with categorises techniques, such as
a general encryption quadtree scrambling, are
introduction to techniques vulnerable to attacks
image into spatial, like the "jigsaw
encryption, frequency, and conundrum" and
providing an hybrid "neighbour ambush."
overview of the domains, Some frameworks,
fundamentals. providing a particularly those
Subsequently, it comprehensive relying on phase spectra,
explores a survey of are incompatible with
comprehensive existing image compression
exploration of methods. standards, limiting their
chaos-based Chaos-based practical applications.
image encryption is Additionally, despite
encryption, highlighted for advancements, chaos-
encompassing its use of based encryption
various methods randomised systems sometimes lack
and approaches key systems sufficient robustness
within this and its against sophisticated
domain. These application in cryptanalysis.
methods include both full and Inconsistencies in their
full encryption selective implementation can
techniques as encryption create security
well as selective frameworks. vulnerabilities.
encryption Full
strategies, encryption
offering insights processes all
into their pixels
principles and uniformly,
applications. The while selective
authors place sig encryption
nificant targets specific
 emphasis on regions for
surveying prior computational
research efficiency.
contributions, Techniques
shedding light such as tent
on noteworthy maps and
developments SCAN-based
within the field.methods in the
Additionally, thespatial domain
paper addresses and phase
emerging spectra
challenges and techniques in
issues that have the frequency
arisen as a domain are
consequence of among the key
these approaches
advancements. discussed.
Hybrid
methods
combining
spatial and
frequency
elements aim
to enhance
security
further
ii A Comprehensive Alyaa I. Today, with the It classifies Techniques relying on
Review of Color Dawood 1, continuous encryption low-dimensional chaotic
increase in the approaches maps may have limited
Image Encryption Qabeela Q.
use of computer based on their key spaces, making
Technology Thabit, Taqwa networks and the algorithms, them vulnerable to
O. Fahad 2023. rapid evolution such as chaotic brute-force attacks.
of information maps, DNA Methods incorporating
technologies. computing, DNA-based or deep
The secure and other learning approaches,
transmission of advanced while robust, often
data over the systems. The suffer from high
Internet has paper computational costs,
become an evaluates these limiting their practicality
urgent necessity methods in for real-time use.
to preserve the terms of their Chaotic map-based
privacy of users security algorithms can be highly
and protect features, sensitive to initial
sensitive efficiency, and conditions, where minor
information resistance to errors may compromise
from theft and attacks. Some security. Furthermore,
distortion. approaches, some methods lack
images are most like chaotic scalability or
of this systems, compatibility with large
transferred data, employ 2D datasets, leading to
so it was and 3D chaotic increased encryption and
necessary to maps for pixel decryption times.
protect it by scrambling
encrypting them and data
using algorithms diffusion,
that ensure the reducing
protection of correlation
information between image
access to the pixels. DNA-
receiver. Color based methods
images contain leverage
sensitive biological
information and principles for
details that must encoding and
be secured and scrambling
protected. This matrices,
paper produces a while deep
comprehensive learning
review of image models are
encryption used to predict
methods and chaotic
classifies them sequences,
based on various enhancing
concepts such as encryption
chaotic maps, strength.
DNA, etc. with Additionally,
comparisons the paper
between existing compares
approaches to these
accessing techniques
 different security using
parameters. parameters
Additionally, the such as key
types of space size,
encryption keys computational
were reviewed efficiency, and
along with some vulnerability
common types to different
of attacks and types of
the most attacks.
important
methods for
measuring
encryption
efficiency.
iii Security of End-to- Saba Inam , The Internet of The research The performance of the
End medical images Shamsa Medical Things introduces a encryption and
encryption system Kanwal, (IoMT) connects deep learning- decryption network is
using trained deep Anousha medical devices based influenced by the quality
learning encryption Anwar, Noor and wearables to encryption and and diversity of the
and decryption Fatima Mirza, enhance decryption training data;
network Hessa Alfraihi healthcare, network insufficient or non-
necessitating designed to representative data may
encryption to safeguard lead to suboptimal
protect sensitive medical results. Additionally, the
patient data. images computational
This study throughout complexity of the deep
proposes deep their learning model can be a
learning-based transmission. concern, potentially
encryption for The requiring significant
medical images methodology resources for training
using involves and deployment. The
Cycle_GAN, a training a deep study also notes that
Generative neural network while the proposed
Adversarial to perform system demonstrates
Network. Cycle both high security, its
GAN enables encryption and resilience against
image decryption advanced cryptographic
transformation tasks, ensuring attacks remains an area
without that the for further investigation.
requiring paired encrypted Moreover, the
training data, images remain generalisability of the
addressing confidential model to different types
challenges in and can only of medical images and
obtaining be accurately its scalability in real-
annotated decrypted by world applications are
medical datasets. authorised aspects that warrant
It uses a entities. The additional research.
generator to study employs
modify images a dataset of
and a medical
discriminator to images,
distinguish real applying the
from fake, with a trained
Binary-Cross network to
Entropy loss assess its
function for effectiveness
training. in maintaining
Experiments on image quality
skin cancer and security.
datasets show
this method
delivers
efficient,
systematic, and
secure
 encryption
compared to
other modern
approaches,
ensuring
confidentiality
and authenticity.
iv Image Cryptography Mr. Shivdatta Image protection The The main limitation or
(Encryption and Pawar, Mrs. is vital as it methodology weakness identified in
Decryption) Surekha Kohle dominates for Image the approach is the
2022. multimedia data. Cryptography efficiency of the
With increasing (Encryption algorithm. Specifically,
digital image and the decryption process is
transmission and Decryption) as significantly slower than
storage, ensuring presented in encryption. While
confidentiality, the journal by encryption takes less
integrity, and Mr. Shivdatta time, decryption is
authenticity has Pawar and roughly four times
become crucial. Mrs. Surekha slower, which can be an
Image Kohle focuses issue in time-sensitive
encryption on applications.
addresses this by transforming Additionally, the
scrambling image data complexity of the
pixels and into a matrix process, involving
reducing pixel format. The matrix manipulations
correlation to process begins like Gaussian
enhance by converting elimination, introduces
security. This the image into computational overhead,
paper reviews a grayscale making it less efficient
various image representation compared to simpler
encryption by averaging encryption algorithms.
 techniques, its RGB
discussing their values.
methods, Gaussian
advantages, and elimination
limitations, with partial
while providing pivoting and
an overview of row exchanges
image is applied to
encryption's this image
significance. matrix,
generating an
upper
triangular
matrix that
serves as the
encrypted
form. The
decryption
process uses
the decryption
key generated
during
encryption to
reverse the
transformation
and retrieve
the original
image.
v Image Encryption Yousef, Ensuring the A Survey of as discussed in the
Algorithms: A Alghamdi, and confidentiality Design and paper, includes the
Survey of Design and ArslanMunir and privacy of Evaluation inherent vulnerabilities
Evaluation Metrics 2024. images shared Metrics" by associated with some
over unsecured Yousef chaotic-based
networks is vital. Alghamdi and encryption methods,
Common Arslan Munir such as the limited range
encryption provides an of chaotic behavior and
methods like extensive susceptibility to chosen-
AES, Twofish, overview of plaintext attacks. For
and RSA, while various image instance, while chaotic
secure, are encryption systems like the Logistic
unsuitable for algorithms and and Arnold maps are
real-time image their fast and effective, they
encryption due corresponding can be prone to reverse-
to their slow evaluation engineering under
speeds and high metrics. The specific conditions,
computational methodology which may compromise
demands. As a adopted in the security.
result, paper includes
researchers have an in-depth
developed classification
specialized of image
algorithms for encryption
image techniques
encryption. This based on the
paper reviews strategies they
and classifies employ, such
these algorithms as chaotic
into seven maps,
approaches permutation-
based on their substitution
techniques. It schemes, and
evaluates each key
 based on management
security, quality, methods. The
and efficiency authors
metrics, evaluate these
providing upper algorithms
and lower based on their
bounds for these security,
evaluations. The computational
discussion also efficiency, and
covers the quality of
strengths and encrypted
weaknesses of images,
various methods making
and their comparisons
suitability for between
different different
applications techniques to
guide future
research and
application in
this field
vi Fast and Secure Rasha S. Ali, The rapid In this article, the paper lacks a
Image Encryption Maha Khalil evolution of we aim to comprehensive security
System Ibrahim, Saad network utilize the analysis. It does not
Using New N. Alsaad technologies, lightweight address potential
Lightweight 2024. where computer Trivium vulnerabilities to
Encryption vision and the algorithm to cryptographic attacks
Algorithm Internet be suitable for like differential analysis,
seamlessly securing chosen-plaintext attacks,
merge, poses transmitted or known-plaintext
significant data on the attacks, the paper does
challenges in networks. Our not provide detailed
safeguarding main information on key
end-user data. contributions management practices,
To address the are as follows: such as key generation,
growing 1. Proposed a distribution, storage, and
need for secure compact sensitivity analysis, the
data in smart synchronous paper does not provide
environments, stream detailed benchmarks or
lightweight and cipher system comparisons with other
efficient security with improved lightweight encryption
solutions are in parameters algorithms.
high that can
demand. offer higher
Notably, the security and
Trivium, a well- better chip size
known encrypt and power
stream cipher consumption
(eSTREAM) performance.
project, has been 2. Outline the
the target guidelines for
of several selecting
attacks. This effective
paper proposes parameters
an innovative when using
approach for fast chaotic
and highly functions.
secured image
encryption
using byte
scrambling and a
modified version
of the
Trivium
algorithm. The
Henon map is
utilized to
generate random
numbers for
permutation,
reducing
time and
increasing
security.
Performance is
measured
on 100 colour
images with
different several
tests. The
results present a
fast and robust
security system
for the
transmission
process. It is
more efficient
than the
traditional
Trivium method.
The encrypted-
decryption
time is reduced
 by 1:24 times,
making it a
quick
security system
for surgical
telepresence and
multimedia
visuals.

vii A Comprehensive N. Mahendiran, The explosive

Analysis on Image C. Deepa 2021. generation of
Encryption data in the
and Compression
digital era
Techniques
directs to the
with the Assessment
requirements of
of Performance
effective
Evaluation Metrics
approaches to
transmit and
store the data.
The necessities
of security and
limited resources
have led to the
development of
image
encryption and
compres sion
approaches,
respectively. The
digital contents
are transmitted
over the internet
which may
subject to
security threats.
To overcome
these limitations,
image
encryption
approaches are
developed.