BCI2001-Data Privacy

Image Steganography Using an Edge Based Embedding Technique

Review 3

REG NO. NNAME

19BCE0053 TARANG GARG

19BDS0159 ADRIJA MUKHOPADHYAY

19BCI0002 PRINCE KUMAR

19BCE0735 ANAND AGARWAL

19BCE2301 DEEP KOTECHA

`
Under the guidance of

Prof. JASMIN T. JOSE

SUMMER SEMESTER SPECIAL 2021-22

 ABSTRACT
Steganography, also known as a carrier or data carrier, is a technique for concealing sensitive
information by encoding it within an audio, video, image, or text file. It is one of the techniques
used to defend sensitive or secret data against nefarious attacks. Both steganography and
cryptography are techniques used to conceal or safeguard secret data. However, they differ in that
steganography conceals the existence of the data, whereas cryptography renders the data
unreadable or hides the meaning of the data.

A tiny piece of information is added to the pixels, scarcely altering the appearance of the image.

The image used for this purpose is known as the cover-image in image steganography, and the
image that results from steganography is known as the stegoimage.

The project's goal is to develop a revolutionary data-hiding method for digital images based on
edge-based algorithms and the LSB (Least Significant Bit) methodology. To further protect user
privacy, standard AES encryption, age and zip code generalization, and username
psedonymization will also be used.

Introduction
A hidden message is included in an image that is conveyed across a network while maintaining its
original appearance. The information can be kept safe and secure in this way. Users can create
their own accounts here and add the shared data—which is first compressed, then encrypted, and
then embedded in a carrier file—by adding it to their accounts. The data is protected by numerous
layers of encryption, and even if an attacker gets to bypass the first layer, the information won't
make any sense because it will only be the carrier file. Even if they are able to recover the data,
there is still another encryption layer to get beyond.

The project's goal is to develop an image embedding method that is reliable, effective, safe, and
privacy-preserving for storing files as images. This method will be based on the LSB approach
and edge-based embedding. To further secure the data, a second layer of security must be applied.
 LITERATURE SURVEY

Authors Algorithms/Methodology Scope for future

S.NO Title of the used work/Limitations
Paper

And year

MUHALIM Least Significant Bit (LSB) In future they will work on

1 INFORMATION Masking and filtering
MOHAME
Transform techniques
HIDING USING D AMIN for finding best approach of
STEGANOGRA SUBARIAH steganography to hide
information.
PHY IBRAHIM
MAZLEEN
(2003)
A SALLEH
MOHD
ROZI
KATMIN

Mstafa, For the Future they are in

2 Information Ramadhan Spatial Domain Watermarking, process of designing a concept
Hiding in Images & Bach, of fingerprinting of images
Using Christian Least Significant bit Watermarking, using watermarking techniques
Steganography so as to introduce more privacy.
Watermarking Frequency Domain Watermarking
Techniques
Discrete Wavelet Transform
(2013)
Discrete Cosine Transform

3 Steganography-A Vinita Pixel value differentiating(PVD) On this method he is willing to

Data Hiding Haridas compromise on some features to
ensure the security of others.
Technique

(2018)

This paper discusses the privacy issue The system under consideration
4 Privacy and Data Shrikant, from an Indian point of view in relation for proposal encompasses all
Protection in Ardhapurkar to difficulties in the legal, technical, areas in three dimensions: legal,
Cyberspace in & and political domains. technical, and political. In the
Indian Srivastava, Especially data in domains: suggested system, numerous
domains have been attempted to
Environment. & Tanu, & Telecommunication be covered while keeping in
Swati, BPO mind the rapidly advancing
(2010) Sharma & e-Commerce state of technology and newly
chaurasiya, Health Sector developing fields. The
 Mr & Vaish, Media suggested approach has given
Abhishek. National Security room for development so that
Tourism new domains can be introduced
without interfering with existing
ones. In order to satisfy both
current needs and future needs,
the suggested system has been
made adaptable and scalable. A
well-designed privacy
framework is crucial not only
for an individual but also for
society and a nation's economic
development.

Nils In the context of big data analysis, they This paper discussed the effects
5 Privacy Issues Gruschka go over the present state of the legal of data protection regulations on
and Data restrictions and analyze various data big data projects and examined
Protection in Big Vasileios protection and privacy-preserving how privacy-preserving
Data: Mavroeidis strategies. strategies might be used by
Additionally, they show and examine utilizing two case studies. The
Kamer two actual research projects that dealt outcomes were very different.
A Case Study
Vishi with sensitive data and actions for Participants in one experiment
Analysis under adhering to the laws governing data were required to provide
GDPR Meiko regulation. They demonstrate the sorts consent in order to address
Jensen of information that pose a privacy risk, privacy concerns around the
(2018) the privacy-preserving strategies that gathering and processing of
were used in line with the law, and biometric data. Additionally,
how these techniques affected the data there were no issues
processing stage and the research encountered throughout the data
findings. analysis stage. Data from an
existing data source were
utilized in the second project. In
this scenario, anonymization of
several data fields was
necessary, which made data
analysis more difficult and
frequently constrained.

6 Steganograph Rosziati The proposed algorithm uses binary In future they will try to
y Algorithm Ibrahim, codes and pixels inside an image. The increase the accuracy of the
to Hide Secret Teoh Suk zipped file is used before it is algorithm.
Message Kuan converted to binary codes to maximize
inside an the storage of data inside the image.
Image Various sizes of data are stored inside
the images and the PSNR (Peak signal-
(2011) to-noise ratio) is also captured for each
of the images tested.

7 Secure QR- 1. Abhijeet i)A three layer layered architecture ii) Threading can be used to reduce
Code Based Mendhe 2. First layer (RSA encryption algorithm) the encryption and decryption
Message Mr. Deepak iii) Second Layer (image time
Sharing Kumar steganography techniques to hide
System Using Gupta 3. Dr. secret message in QR code) iv) Third
Steganograph Krishna Pal Layer (QR Code image is encoded
y Sharma using mask image)
8 Hierarchical 1. Ms. i) LSB method is applied for image Future scope would be towards
Implementatio Moushmee steganography ii) RKO technique is adding user authentication using
n of RKO Kuri 2. Dr. used to split stego image into random password and/or photo.
Technique for Tanuja share and key share at the sender side.
Visual Sarode iii) key share are overlapped using
Steganograph XOR method to form single image or
y stego image iv) reverse image
steganography is applied to get back
the original image

9 Use of Yogita Patel i) Data Encryption using any i) Could be towards adding
Genetic Encryption algorithm ii) LSB public/ private key for
Algorithm and Steganography iii) Genetic algorithm is encryption.
Steganograph used to reshuffle the modified image ii) face recognition facility for
y for Data bits user to introduce more security
Hiding in
image for
Wireless
Network

10 Hiding Data Hesham F. Combining RSA and Huffman coding To increase visual quality and
Using A. Hamed, or DWT storage capacity.
Efficient Aziza I.
Combination Hussein,
of RSA Ashraf A.
Cryptography M. Khalaf,
and Osama,
Compression Fouad Abdel
Steganograph Wahab
y Techniques

11 A Arshiya JPEG (DCT/DWT) algorithms High lossy embedding

comparative Sajid Ansari,
study of Mohammed
recent Sajid
steganography Mohammadi
techniques for ,
multiple Mohammed
image formats Tanvir
[ Parvez

12 Enhanced Jagan Raj, enhanced Least Significant Bit In future they will try to
Least Jayapandiya increase the accuracy of the
Significant Bit n, C. algorithm so that the quality of
Replacement Kavitha, K. the cover image increase.
Algorithm in Sakthivel
Spatial
Domain of
Steganograph
y Using
Character
Sequence
Optimization
13 Hiding Data Abdel The first technique used LSB with no Future work to increase the
in Images Wahab, encryption and no compression. In the accuracy of both the algos . It is
Using Aziza I. second technique, the secret message is clear that they can hide the
Steganograph Hussain, encrypted first then LSB technique is intended data in messages while
y techniques Hesham applied. minimizing its size.
with F.A. Hamed,
compression Hamdy M.
algorithms Kelash,
Ashraf A.M.
Khalaf,
Hanafy M.
Ali

14 An adaptive Vajiheh The implementation results

image Sabeti, indicated that using the IWT
steganography Masomeh Integer Wavelet Transform (IWT) transform with this concept was
method based Sobhani,Sey significantly more successful
on integer ed than using the DWT transform
wavelet Mohammad and that 64 was the appropriate
transform Hossein block size for 512 × 512
using genetic Hashemineja images. The proposed method
algorithm d was successfully embedded in
the image’s edges, and the
stego image quality was
superior to the methods
compared. The test results
indicated that the best selection
sequences for detailed subbands
for embedding were HH, HL,
and LH. The PDH analysis, RS,
and universal steganalysis
attack demonstrated that the
proposed method is sufficiently
secure against steganalysis
attacks.

15 Image Kumar
steganography gaurav, Canny edge detection, dilation operator
based on Umesh and hybrid coding
Canny edge Ghanekar
detection,
dilation
operator and
hybrid coding

16 Information Sanjeev Edge identification plays a

hiding with kumar,Amar Novel fuzzy edge identification relatively crucial role in image
adaptive pal Singh, steganography. Utilization of
steganography Manoj edge locations for embedding
based on Kumar secret data is a simple adaptive
novel fuzzy technique but robust to many
edge attacks thus it is an excellent
identification way of hiding secret
information while preserving a
good perceptibility.
17 A Survey- Shaukat Dar, IBE (Identity Based Encryption), PIR This paper serves as a roadmap
Data Privacy Kamran & (Private Information Retrieval), DES for understanding the threshold
through Javed, Imran (Digital Encryption Standard), and that must be met
Different & Ammar, AES are only a few of the ways that To keep privacy of user data
Methods. Syed & have been addressed in the survey of intact.
Journal of Abbas, Syed various techniques (Advanced
Network & Asghar, Encryption Standards).
Communicati Sohail & These are accompanied by various
ons and Bakar, M & non-conventional encryption
Emerging Shaukat, approaches which relate to Images.
Technologies
Usman.
(2015)

18 Edge-based Saiful Islam, Threshold Selection, The Embedding, The performance of the planned
image Mangat R Extraction by retrieving secret message technique is anticipated to be
steganography Modi & improved if one uses syndrome
, (2014) Phalguni comminttal to writing to cut
Gupta back the number of distortion
that occurred thanks to
embedding.

19 Image Mehdi LSB based Methods, Pixel Value The comparison of existing
Steganograph Hussain, Differnce (PVD) based methods, EMD embedding ways in spacial
y in spatial Ainuddin based methods, MBNS based methods domain was delineate and
domain: A Wahid & PPM based methods highlighted their blessings and
survey Abdul challenges with the sort of
Wahab, tabular and graphical
Yamani Idna illustration.
Bin Idris

20 Chaos based Ratnakirti Canny’s Edge Detection, Payload The planned technique performs
Edge Roy, Scrambling Algorithm, Embedding higher LSBR and PVD in terms
Adaptive Anirban Algorithm of stego image fidelity. Future
Image Sarkar & work can specialise in
Steganomy Suvamoy extending the planned technique
Changder for higher order bits within the
image planes to make amends
for the capability shortage and
conjointly on potential
reduction of the technique’s
time quality.
Overview of the proposed work

Objectives of the Project

The aim of the project is to combine traditional encryption with image steganography and
designing a robust and efficient embedding algorithm. The user credentials have to be
pseudonymized and their details have to be generalized before embedding. Text compression to
reduce the number of bits to be processed and embedded has also been planned.

Software Requirements

• Windows OS/ Mac OS/Linux OS

• Python 3.6 or above

• Python package opencv to read and write image pixels

• numpy==1.15.2
• Pillow==5.3.0
• pycrypto==2.6.1
• PyQt5==5.9.1
• PyQt5-sip==4.19.12
• PyWavelets==1.0.1
• qrcode==6.0
• sip==4.19.8
• six==1.11.0

Hardware Requirements

1 GHz CPU, 4 GB RAM

Architecture Diagram
Embedding Process
➢ Firsty, the message or the plain text and the 128 bit key is taken form theuser through
the GUI built with PyQt5. The key is hashed with Secure Hash Algorithm (sha256).

➢ Now the plain-text is padded and the Random initialization vector is generated by
passing the AES block size. The cipher text is generated byencrypting the text with the
help of the key and the initialization vector through the Cipher Block Chaining Mode.

➢ The ciphered text is further base64 encoded while concatenating theInitialization

vector.
 ➢ Now we take the cover medium as input, convert it to grayscale and find its Discrete
Wavelet Transform (Haar) which we call as the second coefficient. We then get the QR
code version of the ciphered-base64- encoded text with the help of the qrcode
generation library(The Data Correction Level is kept at L). We obtain the Discrete
Wavelet Transformof the QR code again and call it as the first coefficient.

➢ After that we have two coefficients, namely the first coefficient and the second
coefficient. We fuse it together with some values of alpha and beta depending upon the
correction level we want Here we have used 5%and 95% for alpha and beta
respectively.

➢ The fused image is saved as after obtaining its inverse Discrete WaveletTransform
(Haar) as “Stego.png” which can be viewed.

Formula used during embedding/fusing of the images :

Steganographed_Image = aα + bβ

where α = 0.0050 and β=0.9950

and a= QR Image, b = Cover Medium
Decryption Process
➢ The decryption process involves two outcomes. Firstly the QR code isrecovered
and secondly the plain text is recovered.

➢ The user is shown the UI through which they can select the cover medium and the
steganographed image and after that the Discrete wavelettransforms of both the
images will be obtained.

➢ Using the appropriate decryption formula given below, we’ll recover theQR image
and further we’ll recover the encrypted text in the QR code. The encrypted text will
be decrypted using the key provided by the user and the plain text will be recovered
Formula used for Decryption
a = (Steganographed_Image – bβ)/α

where α = 0.0050 and β=0.9950

and a= QR Image, b = Cover Medium

Experimental Results

TOKEN GENERALIZED

Figure 1 : Cover Medium Figure 2 : Steganographed Image

Figure 3 : QR Code from the text

Figure 4 : Coefficient 2 or the DWT transform matrix of the cover medium

Figure 4 : Coefficient 1 or the DWT transform matrix of the QR Code Image

Figure 4 : Steganographed image or the fused image of the cover medium

andthe QR code
Work-around Screenshots

Figure 5 : Running Commands

Figure 6 : GUI Implementation

After we press Encode, a steganographed image is generated which is shown
below stego.png
Figure 7: Fused Image

Figure 8 : QR Recovery
Figure 9: Recovered QR Image

Figure 10 : Process for data Recovery

 Figure 11 : Recovered Message

Evaluation Metrics
Mean Square Error: MSE is the averaged pixel-by-pixel squared difference
between the cover-image and the stego-image.

Where, f & g corresponds to pixel value of cover image and stego image
respectively and m and n represent total number of rows and columns of the
image matrix respectively.

PSNR((Peak Signal to Noise Ratio) : The PSNR is expressed in dB’s and can be
calculated using MSE as

Where, P is the peak signal value of the cover- image, and

 Figure 12: Evaluation Metrics Result

Figure 13: DWT hidden text and LSB hidden text respectively for comparison

Conclusion:
The proposed work is successfully simulated and works as defined whichensures multi-
layered security for transmission throughout the medium.
Comparison is shown which shows a negligible amount of difference in the original image
and the steganographed image. AES being a very good encryption algorithm combined with
text hiding using steganographic algorithms will really provide a new definition to security
and is an area which further can be improved upon. Intensive research work is going on
related to QR code based message/information sharing and to generate a hybrid encryption
algorithm that is robust, efficient, flexible has higher throughput, andhas error correction
capabilities. For overall simulation process, we have utilized the system/machine on the basis
of following parameters. This simulation setup uses the provided packages from the Python
Library to simulate the performance of proposed work. This implementation was thoroughly
tested and was optimized to give the maximum performance for the algorithms. Proposed
System was coded using Python programming language.

Future Work
Further improvements can be made into the compression techniques used to reduce the distortion
and increase capacity. The mapping technique for the pseudonyms can also be further improved.
Efficiency of the embedding algorithm could also be improved to require fewer iterations with a
lower time complexity.
By this, cryptographic and steganographic techniques can be combined with anonymization
techniques to create systems that provide high privacy, security and hide it efficiently in
multimedia files.

References [1] Arshiya Sajid Ansari, Mohammed Sajid Mohammadi, Mohammed Tanvir Parvez,
‘A comparative study of recent steganography techniques for multiple image formats’,
International Journal of Computer Network and Information Security, 2019. [2] Hesham F. A.
Hamed, Aziza I. Hussein, Ashraf A. M. Khalaf, Osama, Fouad Abdel Wahab, ‘Hiding Data Using
Efficient Combination of RSA Cryptography and Compression Steganography Techniques’, IEEE
Access Vol 9, 2021. [3] Jagan Raj, Jayapandiyan, C. Kavitha, K. Sakthivel, ‘Enhanced Least
Significant Bit Replacement Algorithm in Spatial Domain of Steganography Using Character
Sequence Optimization’, IEEE Access Vol 8, 2020. [4] Abdel Wahab, Aziza I. Hussain, Hesham
F.A. Hamed, Hamdy M. Kelash, Ashraf A.M. Khalaf, Hanafy M. Ali, ‘Hiding Data in Images
Using Steganography techniques with compression algorithms’, Telkomnika Vol 17, 2019. [5]
Text book: Cryptography and Network Security Sixth Edition by William Stallings.
REVIEW-1