Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

An Effective Multiple-Bit Data Abstraction and

Lossless Retrieval without Shifting
P. Arunkumar1; S. Divya2; E. Baskar3; R. Brindha Devi4
Student1; Student2; Student3; Assistant Professor 4
Department of Computer Science and Engineering Sri Venkateswaraa College of Technology

Abstract:- Cryptography and Steganography are two I. INTRODUCTION

major forms of enforcing data privacy, with the former
being prevalent. For small payloads, Steganography is Image Steganography is a technique under the domain of
preferred, as the data remains concealed and abstracted. information security, which works by concealing information
The process involves embedding bits onto a carrier image. or mission critical data within images that meet certain
However, the disadvantage of steganography is that the requirements.
carrier image gets distorted, which is a major concern when
it comes to pixel sensitive images, such as forensic and This technique is also used to protect secret or sensitive
medical images that are mission critical. Furthermore, due data from malicious attacks by embedding the data into images
to modifications in the pixel values, the images get distorted, and retrieving the same. Hiding sensitive data is easily
which can be irrecoverable. Hence, the technique of achievable for ordinary images. However, for embedding data
Reversible Data Hiding (RDH) can be used as a solution. on sensitive images, such as medical, forensic or military
Usually two methods are used under RDH, namely images, the image tends to get distorted due to manipulation of
Histogram Shifting (HS) and Difference Expansion (DE). pixel values, and this distortion becomes permanent, leading to
poor quality of images. Hence, reversible data hiding schemes
Histogram Shifting method shifts all pixel values are preferred. There are various methods for achieving RDH.
between pairs of peak and zero points, which allows data to
be embedded in the intermediate points. However, this One of the most popular Reversible Data Hiding technique
method also depends on frequencies of pixel values on the is Difference Expansion (DE), which is simple in terms of
image, and the embedding capacity varies greatly on the implementation. As the name suggests, this RDH scheme looks
image. In Difference Expansion Method, the difference for expandable differences to embed pixels into.
between a pixel and its adjacent pixels is considered to
embed a single bit intoone of the pixels. Here, a scheme to The existing DE techniques execute one layer embedding
embed multiple bits in a pixel is proposed, which greatly in a difference image and use the histogram shifting method.
enhances the embedding capacity. Initially, three The evident disadvantage of this technique is that the image
horizontally adjacent pixels are taken and the difference quality could be severely degraded as more layers are
between the rightmost and leftmost pixel iscalculated. embedded. If less layers have to be incorporated, the embedding
capacity is low. Moreover, computing and expanding the
This difference is used to classify the centre pixel as difference involves complex computation.
embeddable or non-embeddable. If the pixel is embeddable,
anew pixel value is calculated, which comprises of the secret In this project, a reversible data hiding scheme that can
bits, offset bits and flag bits, without the use of any embedmultiple bits in a pixel without shifting is discussed as a
difference expansion or histogram shifting process. The solution for the above-mentioned disadvantage.
secret bits and the original pixel value can be easily
recovered from the modified value using the offset and flag This scheme specifies clear criteria for the embeddability
bits. This scheme has higher embedding capacity, and lower of a pixel and the same criteria can be used to retrieve during
complexity in terms of computations, when comparedwith extraction of the data from the embeddable pixel. Flag bits,
other RDH schemes. offset bits and 2 embedding bits are computed.

Keywords:- Information Security, Steganography, Reversible

Data Hiding, Histogram Shifting, Difference Expansion,Carrier
Image.

IJISRT24MAR1334 www.ijisrt.com 1645

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

II. PROPOSED ARCHITECTURE The next three bits and two bits denote the offset andflag
bits respectively. If the flag bits F(i, 2j) = 11 and the offset Q(i,
 The proposed scheme consists of two algorithms, namely:- 2j) = T, then no secret can be stored and the first three bits are
 Embedding algorithm Retrieval Algorithm occupied by another offset Q1(i, 2j) , which is calculated as
shown in the following equation.
A. Embedding Algorithm
The embedding algorithm makes use of correlation of Fig 1.2 Binary format of new pixel value when F=11 and
three horizontally pixels at a time, to conceal multiple bits. Offset = 2n – 1
Considering a grayscale image, I with width w and height h
pixels, the possible values for each pixel in theimage would lie Finally, the new pixel value is calculated by: -
between 0 to 255. For i ∈ [1, h] and j ∈[1, ⌊w/2⌋], the difference 7
d for any pixel p(i,2j) isgiven by : - 𝑝′(𝑖,2𝑗) = Σ𝑎𝑖×2𝑖
𝑖=0
For a threshold T, the embeddability e (i, j) is determined
asper the following equation: -
B. Retrieval Algorithm
If e(i, 2j) is 1, then the point is classified as an embeddable Now, for the received stego image I and for a threshold T,
point (EP). If e(i,2j) is 0, then the point is a non-embeddable the reverse process of the embedding algorithm can be
point (NEP). A NEP cannot be used to embed any secret, as performed to extract the secret bits and the carrier image I′,
there is very correlation between three adjacent pixels. where that p(i,2j) is a stego pixel in I and p′(i, 2j) is the
corresponding pixel in retrieved image I′.
Then the EMAX (i, 2j) and EMIN (i, 2j) of p(i,2j), which
denotes the local maximum and minimum respectively, is given III. RELATED WORK
by the following equations: -
Reversible Data Hiding is a prevalent task in Image
Hence, it’s local offset OS(i, 2j) is calculated by: - Steganography. There are various Image Steganography
methodologies that are prevalent to achieve Reversible Data
Now, a new value p’ (i, 2j) consisting of secret to be Hiding (RDH).
embedded, offset bits and flag bits is calculated. The newvalue is
constructed in binary format as shown in Figure 1.1. The EP One of the most popular techniques for RDH is Difference
p(i,2j) is then replaced by the calculated value. It’s two flag bits, Expansion (DE), which is simple to implement. In one such
denoted by F(i,2j) can be calculated by: - scheme, an image is split into multiple blocks of pixels, eachof
size 2 × 2, and each block is converted into a vector by
Also, T must hold the following condition: - Where n is calculating the difference between the pixel values.
the length of offset (in bits) in p’(i,2j).
A set of equations, depending on the algorithm, is used to
The offset Q(i,2j) can be calculated as per the following: manipulate the pixel values. Another set of equations, which are
an inverse of the former set, is used to retrieve the secret and
Where Dec2Bin() is a procedure that converts a decimal old pixel value. However, in this method, the location map
number to its binary equivalent. Depending on the flag, local becomes significantly high.
offset bits, local maximum and minimum, p’(i, 2j) can be
calculated as per the following figure: - A. Merits:

Fig 1.1 Binary format of new pixel value  One more technique that is commonly used is Histogram
Shifting (HS). In this technique, two types of pixel values,
If F = 00, 01 or 10 then the first three bits denote the secret, namely zero point and peak point arefound. Zero point is the
the middle three bits denotethe offset Q(i,2j) which is calculated pixel value which appears theminimum number of times in
using Eq.19 , and the last two bits denote the flag bits F(i,2j), the image, and Peak Point is the value, which appears
which is calculated by Eq. 17. maximum in the image. Then, a pixel with zero-point value
and a pixel with peak value is taken together to hide the
If the flag bits F(i,2j) = 11 and the offset Q(i,2j) is not T, secret.
then three secret bits can be embedded as three most significant  The value of pixels in between the pair is increased by 1
bits of p’(i,2j). unit. In other words, the values of pixels between the zero
and peak points in the histogram are shifted by 1 unit.
However, in this method, the embedding capacity varies
greatly as it depends on the number of peak and zero points,

IJISRT24MAR1334 www.ijisrt.com 1646

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

and hence the payload is also low B. Demerits:

 The proposed model has a higher embedding capacityeven
though the image contains evenly distributed pixel values as  After generating the stego image it was observed that the
modified pixels in those regions can be easily noticed. quality of the embedded image is notably lower with respect
 Further the embedded text and original image can beeasily to the carrier image.
retrieved with good quality.  After performing the image retrieval, the quality of the
 The problem of underflow is completely avoided in this retrieved image was of good quality, but some of the pixels
method because the stego pixel is directly constructed using got modified.
the adjacent pixels without performing HS or DE schemes.  The algorithm did not account for the overflow of pixel
values.
Multiple bits are embedded at the same time by altering the
pixel values between its neighbours, which increases the
embedding capability.

IV. RESULTS FOR BICYCLE.BMP

Fig 1: Original Image for Bicycle.bmp

Fig 2: Embedded Image for Bicycle.bmp

Integration with third-party APIs enriches the system with real-time data and services, including travel booking.

IJISRT24MAR1334 www.ijisrt.com 1647

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

Fig 3: Retrieved Image for Bicycle.bmp

V. RESULTS FOR AEROPLANE.BMP

Fig 4: Original Image for Aeroplane.bmp

IJISRT24MAR1334 www.ijisrt.com 1648

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

Fig 5: Embedded Image for Aeroplane.bmp

Fig 6: Retrieved Image for Aeroplane.bmp

IJISRT24MAR1334 www.ijisrt.com 1649

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

VI. RESULTS

The base paper has been implemented based on the

architecture proposed and the results obtained have been
tabulated in table 1. To analyse the effectiveness of the
proposed system, three performance metrics were evaluatedviz.
Embedding Capacity, Payload and PSNR.
The following metrics were evaluated for 6 grayscale
The equation for calculating PSNR values between two 512x512 test images as tabulated in table 1.
images is given as follows:

Table 1: Analysis of Images using EC, Payload and PSNR

VII. CONCLUSION It eliminates any need to compute the difference in

histogram calculation and thus reduces computational
In this project, an efficient method for embedding multiple complexity significantly. It also avoids involving multiple
bits in a pixel is proposed for reversible data hiding that doesnot layers of pixelsto notably simplify the data-hiding process.
employ any shifting.
The proposed multiple bit reversible data hiding scheme
By exploiting the relationship among its two horizontally without shifting works best with the threshold value of 7
adjacent neighbours, multiple bits can be directly embeddedin yielding the highest PSNR vs. Payload ratio, as per the plot. The
an EP with a strong correlation with its horizontal neighbours, proposed scheme also yields good quality of the retrieved cover
without modifying any NEP. image and facilitates lossless retrieval of the embedded data.

Fig 7: 1PSNRvs.Payloadplot

IJISRT24MAR1334 www.ijisrt.com 1650

Volume 9, Issue 3, March – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24MAR1334

Furthermore, the PSNR vs. Payload graph was plottedto

analyse the trend for different threshold values to maximize the
embedding capacity and the visual quality of the stego-image.

FUTURE WORKS

The proposed algorithm has further scope for

improvement in terms of the quality of the embedded image and
the error-free retrieval of the embedded secret. A little tweaking
to the algorithm can yield better quality of the stegoimage and
minimal distortion to the carrier image. The embedding
capacity can be improved by adding more criteria for deciding
the embeddability of a pixel apart from the offset and flag bits
can be established. This also facilitates the error free embedding
and retrieval of the data. We plan to improve the embedding
algorithm by handling a few more cases while deciding the
embeddability of a pixel and introduce a pixel map to keep track
of the overflow bits during embedding. The security for the
cover image can be enhanced by the using a suitable encryption
algorithm. Instead of directly embedding the data into the
carrier image we first encrypt the carrier image and embed data
into the resulting image. During the retrieval, the user can
choose to retrieve only the image, or only the secret, or both
communities and cultures. Through curated recommendations
and educational resources, TourTrove inspires travelers to
engage in mindful and conscientious travel experiences that
leave a positive imprint on the world.

REFERENCES

[1]. Wang, W. (2020). An efficient multiple-bit reversible

data hiding scheme without shifting. Multimedia Tools
and Applications, 79(1), 555-579.
[2]. Arham A, Nugroho HA, Adji TB (2017) Multiple layer
data hiding scheme based on difference expansion of
quad. Signal Process 137:52–62
[3]. Wang J et al (2017) Rate and Distortion Optimization for
Reversible Data Hiding Using Multiple Histogram
Shifting. IEEE Transactions on Cybernetics 47(2):315–
326

IJISRT24MAR1334 www.ijisrt.com 1651