IMAGE STEGANOGRAPHY

Progress Report

BACHELOR OF TECHNOLOGY
in

COMPUTER SCIENCE & ENGINEERING

by
DHRUV GOEL AKHILSH PARTAP
SINGH
Enrollment No: Enrollment
44114802717 No:44514802717

Guided by

Ms. NEELAM SHARMA

(ASSISTANT PROFFESOR, MAIT)

DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING

MAHARAJA AGRASEN INSTITUTE OF TECHNOLOGY
(AFFILIATED TO GURU GOBIND SINGH INDRAPRASTHA
UNIVERSITY)
1.INTRODUCTION

The growing use of Internet needs to take attention while we send and receive
personal information in a secured manner. For this, there are many approaches
that can transfer the data into different forms so that their resultant data can be
understood if it can be returned back into its original form. This technique is
known as encryption. However, a major disadvantage of this method is that the
existence of data is not hidden. If someone gives enough time then the
unreadable encrypted data may be converted into its original form.
A solution to this problem has already been achieved by using a “steganography”
technique to hide data in a cover media so that other cannot notice it. The
characteristics of the cover media depends on the amount of data that can be
hidden, the perceptibility of the message and its robustness.

1.1 STEGANOGRAPHY VS CRYPTOGRAPHY

Basically, the purpose of cryptography and steganography is to provide secret

communication. However, steganography is not the same as cryptography.
Cryptography hides the contents of a secret message from a malicious people, whereas
steganography even conceals the existence of the message. Steganography must not be
confused with cryptography, where we transform the message so as to make its meaning
obscure to malicious people who intercept it. Therefore, the definition of breaking the
system is different. In cryptography, the system is broken when the attacker can read
the secret message. Breaking a steganographic system need the attacker to
detect that steganography has been used and he is able to read the embedded message.
In cryptography, the structure of a message is scrambled to make it meaningless and
unintelligible unless the decryption key is available. It makes no attempt to disguise or
hide the encoded message. Basically, cryptography offers the ability of transmitting
information between persons in a way that prevents a third party from reading it.
Cryptography can also provide authentication for verifying the identity of someone or
something.
In contrast, steganography does not alter the structure of the secret message, but hides
it inside a cover-image so it cannot be seen. A message in cipher text, for instance, might
arouse suspicion on the part of the recipient while an “invisible” message created with
steganographic methods will not. In other word, steganography prevents an unintended
recipient from suspecting that the data exists. In addition, the security of classical
steganography system relies on secrecy of the data encoding system. Once the encoding
system is known, the steganography system is defeated.

2.PROBLEM STATEMENT

This project addresses the security problem of transmitting the data over internet
network, the main idea coming when we start asking that how can we send a
message secretly to the destination? The science of steganography answers this
question. Using steganography, information can be hidden in carriers such as
images, audio files, text files, videos and data transmissions.
In this document, we proposed some methods and algorithms of an image
steganography system to hide a digital text of a secret message.

3.PROJECT SCOPE

Our project scope is developed for hiding information in any image file to ensure
the safety of exchange the data between different parties and provide better
security during message transmission.
The scope of the project is implementation of steganography tools for hiding
information includes any type of information file and image files and the path
where the user wants to save image and extruded file. We will use LSB
technique; the proposed approach is to use the suitable algorithm for embedding
the data in an image files; we will show a brief of this algorithm that we used to
hiding data.

4.TECHNOLOGY STACK

 MATLAB
 DATA ACQUISITION TOOLBOX
  IMAGE PROCESSING TOOLBOX
 IMAGE ACQUISITION TOOLBOX

5.ARCHITECTURE

Steganography is the art of hiding and transmitting data through apparently

innocuous carriers in an effort to conceal the existence of the data, the word
Steganography literally means covered or hiding writing as derived from Greek.
Steganography has its place in security. It is not intended to replace cryptography
but supplement it.
Hiding a message with Steganography methods reduces the chance of a message
being detected. If the message is also encrypted then it provides another layer of
protection. Therefore, some Steganographic methods combine traditional
Cryptography with Steganography; the sender encrypts the secret message prior
to the overall communication process, as it is more difficult for an attacker to
detect embedded cipher text in a cover. Hidden information in the cover data is
known as the "embedded" data and information hiding is a general term
encompassing many sub disciplines, is a term around a wide range of problems
beyond that of embedding message in content. The term hiding here can refer to
either making the information undetectable or keeping the existence of the
information secret. Information hiding is a technique of hiding secret using
redundant cover data such as images, audios, movies, documents, etc.

Figure 5.1: steganography architecture

6.WORK DONE

6.1 STUDY OF DIFFERENT TYPES OF STEGANOGRAPHY

In General, Steganography is classified into four types as follows:
 Image steganography: It is a process of concealing the secret message inside
the cover image in such a way that the existence of the secret message is
disappeared and the cover image seems to be original.
 Audio steganography: Digital sound files are used to hide a secret message by
vaguely changing the binary sequence of a sound file, which is known as audio
steganography.
 Video steganography: Video files can be defined as a collection of images and
sounds combined together, thus, most of the introduced images and audio can
be used and applied to the digital video files. In fact, large amount of secret data
can be embedded inside the video files, since the video file is a moving stream
of images and sounds.
 Text steganography: Text steganography basically refers to the information
that is hidden in text files. The text steganography includes everything from
manipulating and changing text formatting, word changing within the text,
producing and generating random sequences or using context-free language
grammars to generate readable texts.

For our project we have worked on image steganography out of all the 4 types.

6.2 STUDY AND ANALYSIS OF VARIOUS TECHNIQUES OF

IMAGE STEGANOGRAPHY
There exists a large variety of image steganography techniques. Some of these
techniques are more complex than the others, and all of them have respective
strong and weak points. Image steganography techniques can be classified into
spatial domain (image domain) steganography, transform domain (frequency
domain) steganography, spread spectrum steganography and model-based
steganography. Figure 6.1 shows a classification tree of image steganography
techniques.
 Figure 6.1: different techniques of image steganography

6.2.1 SPATIAL DOMAIN METHODS

spatial domain Steganography technique refers to methods in which data hiding
is performed directly on the pixel value of cover image in such a way that the
effect of message is not visible on the cover image. The spatial domain methods
are classified as following:
i) LEAST SIGNIFICANT BIT (LSB): - LSB is one the technique of spatial
domain methods. LSB is the simple but susceptible to lossy compression
and image manipulations. Some bits are change directly in the image pixel
values in hiding the data. The technique works by replacing some of the
information in a given pixel with information from the data in the image.
Normally, An LSB algorithm replaces the most-right bits of a cover files
bytes. In case a bit of the cover image C(i, j) is equal to the bit of a secret
massage (SM) that to be embedded, C(i,j) stay untouched, otherwise C(i, j)
is set to bit of a secret massage (SM). The LSB method is characterized by
simplicity, high capacity, easy to understand and implement, and can't be
noticed by the naked eye. However, the drawbacks of this methods are that
lacks robustness (Easy manipulation by attackers), susceptible to noise,
scaling and cropping. Eg.
 Figure 6.2: example of LSB conversion

ii) PIXEL VALUE DIFFERENCING (PVD): - In 2003, Da-Chun developed

a new embedding method, called Pixel Value Differencing (PVD), based on
the difference between pixel values. First, they divide the cover image into
non-overlapping blocks having two connecting pixels. Then, they modify
each block difference. They found that the larger the difference into original
pixel value, the greater the modification will be. They also found that the
embedded secret bits number depends on the pixel case that will be in
smooth area or in edge area. In smooth area, the difference between adjacent
pixels is less whereas in edge area it is more. Therefore, the data that is
embedded into edge area pixels is more than embedded into smooth area.

iii) GRAY LEVEL MODIFICATION (GLM): -In 2004, Potdar discussed a

technique based on a mathematical function. This technique maps data by
altering gray levels of the pixels without embedding or hiding it and uses
the conception of even and odd numbers in mapping the data in the cover
image. For example, even values are mapped with zero and odd values are
mapped with one. The gray level modification method is characterized by
low computational complexity and high capacity.

6.2.2 TRANSFORM DOMAIN METHODS

Transform Domain applies image transformation and manipulation of
algorithm. In transform domain steganography, embedding the secret requires
transforming the image from the spatial domain to the frequency domain by
using any of the transforms, for example, Discrete Cosine Transform (DCT),
Discrete Wavelet Transform (DWT), and Single Value Decomposition (SVD).
After the transformation process, the embedding process will be done in proper
transform coefficients.
i) DISCRETE COSINE TRANSFORM (DCT): - Discrete Cosine
Transform (DCT) is based on transforming signal or image from spatial
domain to frequency domain. The DCT split the image as shown in
figure 6.3 up to spectral sub-bands (parts) of different significance with
respect to the visual quality of the image. Embedding positions Choices:
(i) Low frequency coefficients: Bad invisibility, because human eye is
sensitive to noise on it, as it contains the image visual parts, (ii) High-
frequency coefficients: bad robustness, as the image could be corrupted
through noise attacks or compression, and (iii) Middle-frequency
coefficients: good invisibility and robustness, so it is the best choice.

Figure 6.3: - discrete cosine transform frequency bands

Image is broken into 8×8 pixel blocks and transforms these pixel blocks
into 64 DCT. Working from left to right, up to down, the DCT is applied
to each block. Through quantization table each block is compressed to
scale the DCT coefficients and message is embedded in DCT
coefficients. The array of compressed blocks that constitute the image is
stored in drastically reduced the amount of space. When desired, image
is reconstructed through decompression, a process that uses the Inverse
discrete cosine transform i.e., IDCT is applied.

Figure 6.4: - discrete cosine transform

ii) DISCRETE WAVELET TRANSFORM (DWT): - It is used to
transform the image from a spatial domain to the frequency domain. In
the process of steganography DWT identifies the high frequency and
low frequency information of each pixel of the image. It is mathematical
tool for decomposing an image hierarchically. It is mainly used for
processing of non-stationary signals. The wavelet transform is based on
small waves, Known as wavelets, of different frequency and limited
duration. It provides both frequency and spatial description of the image.
Wavelets are created by translations and dilations of a fixed function are
known as mother wavelet. DWT performs in one dimension and in the
two-dimensional plane. The DWT is the accurate model than the DFT
or the DCT and it is multi resolution description of the image. The
current image compression standard JPEG 2000 is based on the wavelet
transforms.

iii) SINGULAR VALUE DECOMPOSITION (SVD): - Singular Value

Decomposition (SVD) is a mathematical technique based on a linear
algebra theorem which declares that the rectangular matrix (A) can be
analyzed into three matrices: U (Orthogonal matrix), S (Diagonal
matrix), and V (Transpose of an orthogonal matrix). The theorem is
presented usually like: 𝐴 = 𝑈𝑆𝑉^𝑇. The SVD is characterized by
resultant matrices size from SVD is not fixed (square or rectangle) and
Image singular values (SVs) which preserve image most energy, resist
against attacks and having intrinsic algebraic image properties.
However, the drawback of this method is diagonal that will appear in
the extracted watermark.

6.2.3 SPREAD SPECTRUM

The concept of spread spectrum is used in this technique. In this method the
secret data is spread over a wide frequency bandwidth. The ratio of signal to
noise in every frequency band must be so small that it becomes difficult to detect
the presence of data. Even if parts of data are removed from several bands, there
would be still enough information is present in other bands to recover the data.
Thus, it is difficult to remove the data completely without entirely destroying
the cover. It is a very robust approach used in military communication.

6.2.4 MASKING AND FILTERING

This approach is used to hides the data by marking an image. This approach is
valuable where watermarks become a portion of the image. The data will be
embedded where the more significant part of the image rather than hiding it into
the noisy portion. The watermarking techniques are more integrated into the
image and it can be applied without the fear of destruction of the image. This
technique is used in 24 bit and grey scale images.

6.2.5 MODEL BASED STEGANOGRAPHY

Model based steganography is a model which is presented in order to overcome
the weaknesses in the data hiding embedding process such as: problems that
faces spatial domain, Stego image distortions and modifications. This model is
considered as statistics aware steganography or adaptive steganography. This
new method helps to embed secret message taking care of overcoming previous
drawbacks. In this method, the embedding process works as follows. The cover
image was divided into two parts: part that was not altered during embedding
and the other part were used to carry secret message without modifying the
cover statistical properties. The hidden message is supposed to be a random
uniform stream of bits. By using an entropy decoder (which is chosen according
to probability conditional distribution), the hidden message was processed. For
extracting, first, entropy encoder was used. Then, the stego message was
separated into two parts. After that, Probability distribution was calculated,
finally, the secret message is obtained.

6.2.6 ANALYSIS OF VARIOUS TECHNIQUES

6.2.6.1 PARAMETERS CONSIDERED

i) ROBUSTNESS: - Robustness refers to the ability of embedded
data to remain intact if the stego- image undergoes
transformations, such as linear and non-linear filtering,
 sharpening or blurring, addition of random noise, rotations and
scaling, cropping or decimation, lossy compression.

ii) PEAK SIGNAL TO NOISE RATIO (PSNR): - The image

steganography system must embed the content of hidden
information in the image so that the quality of the image should
not change. PSNR is commonly used to measure the quality of
reconstruction of lossy compression techniques Larger the
PSNR value indicates the better quality of image i.e., less
distortion. PSNR is the ratio of the maximum signal to noise in
the stego image.

iii) MEAN SQUARE ERROR (MSE): - MSE is another

watermarked image measurement of Quality. In the ideal case,
Mean Square Error should be zero. In real, this is can't be
achieved to the watermarked image, minor MSE is good.

iv) CAPACITY: - Maximum size of information can be embedded

into image.

v) COMPLEXITY: - How much expensive it is computationally

for embedding and extracting a hidden message?

6.2.6.2 RESULTS

TECHNIQUES DOMAIN INVISIBILITY CAPACITY ROBUSTNESS COMPLEXITY

LSB SPATIAL HIGH HIGH MEDIUM LOW

SPREAD SPECTRUM SPATIAL HIGH LOW MEDIUM MEDIUM

DCT TRANSFORM HIGH MEDIUM MEDIUM MEDIUM

DWT TRANSFORM HIGH MEDIUM HIGH HIGH

PVD SPATIAL HIGH MEDIUM LOW LOW

SVD TRANSFORM HIGH MEDIUM MEDIUM MEDIUM

Table 6.1: comparison of various techniques of steganography

 Keeping in mind the above results we have selected LEAST SIGNIFICANT
BIT (LSB) for our project of image steganography due to low complexity, high
capacity and medium PSNR.

7. NEXT TARGET
 Applying LSB for hiding text in image using MATLAB.

 Applying LSB for hiding image in image using MATLAB.

 Reducing noise in cover image so as to increase PSNR for more secure

communication using DWT in MATLAB.

 Try to mix 2 techniques of LSB and DWT for more secure communication

8. CONCLUSIONS
Although only some of the main image steganographic techniques were
discussed in this document, one can see that there exists a large selection of
approaches to hiding information in images. All the major image file formats
have different methods of hiding messages, with different strong and weak
points respectively. Where one technique lacks in payload capacity, the other
lacks in robustness. For example, the patchwork approach has a very high level
of robustness against most type of attacks, but can hide only a very small amount
of information. The proposed approach in this project uses a new steganographic
approach called image steganography. The application creates a stego image in
which the personal data is embedded inside the cover file image. The Least
Significant Bit algorithm in this project for developing the application which is
faster and reliable and compression ratio is moderate compared to other
algorithms.
REFERENCES
1. G. Kaur and A. Kochhar, “A Steganography Implementation based on LSB & DCT,” vol. 4, no.
1, pp. 35–41, (2012)
2. P. Chen and H. Lin, “A DWT Based Approach for Image Steganography,” Int. J. of Applied
Sci. Eng., vol. 4, no. 3, pp. 275–290, (2006)
3. J. Gupta, “A Review on Steganography techniques and methods,” vol. 1, no. 1, pp. 1–4, (2015)
4. H. Kaur and J. Rani, “A Survey on Different Techniques of Steganography,” MATEC Web of
Conferences, ICAET, (2016)
5. A. Fkirin, G. Attiya and A. El-Sayed, “Steganography Literature Survey, Classification and
Comparative Study,” vol. 5, no. 10, pp. 1-10, (2016)
6. T. Bhattacharya, N. Dey, and S. R. B. Chaudhuri, “A Session based Multiple Image Hiding
Technique using DWT and DCT,” International Journal of Computer Applications, vol. 38, no.
5, pp. 18–21, (2012)
7. M. S. Subhedar and V. H. Mankar, “Current status and key issues in image steganography: A
survey,” Computer Science Review, vol. 13–14, pp. 95–113, (2014)
8. D. C. Wu and W. H. Tsai, “A steganographic method for images by pixel-value differencing,”
Pattern Recognition Letters, vol. 24, no. 9–10, pp. 1613–1626, (2003)
9. E. E. A. Elgabar and H. A. A. Alamin, “Comparison of LSB Steganography in GIF and BMP
Images,” International Journal of Soft Computing and Engineering (IJSCE), vol. 3, no. 4, pp.
79–83, (2013)
10. L. K. Saini and V. Shrivastava, “A New Hybrid DWT-DCT Algorithm for Digital Image
Watermarking,” International Journal of Advance Engineering and Research Development
(IJAERD), vol. 1, no. 5, pp. 1–8, (2014)
11. Dipti Kapoor Sarmah and Neha Bajpai, “Proposed System for Data Hiding using Cryptography
and Steganography”, International Journal of Computer Applications, Vol. 8, No. 9, pp. 640-
647, (2010)
12. Hemang A. Prajapati and Nehal G. Chitaliya, “Secured and Robust Dual Image Steganography:
A Survey”, International Journal of Innovative Research in Computer and Communication
Engineering, Vol. 3, No. 1, pp. 534- 542, (2015)