Security and imperceptibility improving of image steganography using pixel allocation and random function techniques

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 12, No. 1, February 2022, pp. 694~705
ISSN: 2088-8708, DOI: 10.11591/ijece.v12i1.pp694-705  694
Journal homepage: http://ijece.iaescore.com
Security and imperceptibility improving of image
steganography using pixel allocation and random function
techniques
Noor Alhuda F. Abbas1
, Nida Abdulredha1
, Raed Khalid Ibrahim2
, Adnan Hussein Ali3
1
Department of Technique Computer Engineering, Al-Esraa University College, Baghdad, Iraq
2
Department of Medical Instruments Engineering Techniques, Al-Farahidi University, Baghdad, Iraq
3
Department of Information Communication and Technology, Institute of Technology Baghdad, Middle Technical University,
Baghdad, Iraq
Article Info ABSTRACT
Article history:
Received Sep 6, 2020
Revised Jul 15, 2021
Accepted Jul 28, 2021
Information security is one of the main aspects of processes and
methodologies in the technical age of information and communication. The
security of information should be a key priority in the secret exchange of
information between two parties. In order to ensure the security of
information, there are some strategies that are used, and they include
steganography and cryptography. An effective digital image-steganographic
method based on odd/even pixel allocation and random function to increase
the security and imperceptibility has been improved. This lately developed
outline has been verified for increasing the security and imperceptibility to
determine the existent problems. Huffman coding has been used to modify
secret data prior embedding stage; this modified equivalent secret data that
prevent the secret data from attackers to increase the secret data capacities.
The main objective of our scheme is to boost the peak-signal-to-noise-ratio
(PSNR) of the stego cover and stop against any attack. The size of the secret
data also increases. The results confirm good PSNR values in addition of these
findings confirmed the proposed method eligibility.
Keywords:
Digital image-steganographic
Huffman coding
Odd/even pixels' allocation
PSNR
Random map function
This is an open access article under the CC BY-SA license.
Corresponding Author:
Adnan Hussein Ali
Department of Information Communication and Technology, Institute of Technology Baghdad Middle
Technical University
Baghdad, Iraq
Email: dr.adnan@mtu.edu.iq
1. INTRODUCTION
The mention of steganography for the science that converts a message into a form that certifies a
complete inability to discover any information hidden in the carrier. A system of steganography warrants each
secret information can be kept undetectable [1]. The word of steganographic originates from a Greek term that
means as a protected writing. It is regarded as a unique area of data concealment, also considered as an art of
science for transmission that is not visible. The aim of making the communication invisible in order to secret
data hiding inside the cover image (CI), thereby enhancing its imperceptibility. The existence of the secret
information (SI) is often recognized by just the sender's and receiver's [2].
The steganography elements, basically included an information, cover object, a stego key for
improved security and embedding mechanism. The carrier object in which the SI is hidden in could be a video,
an audio, text or image. Steganography is being utilized in various applications effectively, but it often can be
very dangerous so it could be used by attackers for sending trojans and viruses in order to manipulate sensitive

Int J Elec & Comp Eng ISSN: 2088-8708 
Security and imperceptibility improving of image steganography … (Noor Alhuda F. Abbas)
695
systems. More so, with the use of this information-hiding technology, criminals or terrorists may be enabled to
exchange secrete information [3]. The payload is described as the size of secret data which can be hidden within
the cover object successfully without creating visible artifacts in stego images. In order to measure payload,
bits per pixel (bpp) is used. If each pixel is used in hiding 1 bit of data, then the payload of a steganography
algorithm will be 1 bpp or 12.5%.
The word robustness refers to the steganography algorithm's durability against different forms of
statistical and basic attacks in steganography. If it is not possible to easily alter or retrieve the data that is
concealed in a cover image using image processing operations, the steganography algorithm used is considered
robust. Scaling, cropping, image rotation, and noise are some of the image processing activities that can be
used in the alteration or retrieval of concealed messages. The issue of robustness is crucial in when
watermarking techniques are used because of copyright protection [4]. The notion of imperceptibility is
synonymous with undectability, which can be calculated using various image quality measurement metrics,
such as peak-signal-to-noise-ratio (PSNR) [5], [6].
The correlation between hiding payload capacity (PC) and quality of stego media (SM) is purely
described via a balance which authors attempt to achieve. Most often, the quality of stego medium is reduced
by concealment of huge amounts of data within a cover medium. Consequently, hiding capacities have
continued to remain comparatively low because of this negative effect on the quality of stego [7]. With respect
to mechanism for embedding, the techniques of steganography are partitioned to two important classes, which
are spatial-domain (SD) and transform-domain (TD). Regardless of the advantages provided by current
methods, problems such as: i) the use of inadequate embedding algorithms may generate visually distorted
stego images (SI), which in turn increases the likelihood of human visual system detection, and
ii) imbalance between image quality, computational complexity, capacity for payload and security; thus,
making them inappropriate. This study, the use of our mechanism alongside with chaotic method is employed
in developing an efficient scheme in spatial domain, with the aim of addressing the aforementioned problems.
This paper makes the following key contributions: i) introduce an effective digital image steganography with
achieving a good quality of image, payload and security; ii) identify the random pixel used to embed hidden
information, while the random function is used to boost the system's resilience against trackers' attempts to
discover which pixel to embed; iii) the embedding of the secret information is done in a random region within
an image through the use of the spatial-domain (SD) of the cover-image (CI) using the odd/even pixel
allocation. This way, the quality of the stego-images is boosted while an extraction of the secret data is made
difficult
2. PRELIMINARIES
2.1. Least-significant-bit LSB substitution
The representation of least-significant-bit (LSB) considers easy and conventional process applied for
inserting secret data enclosed by covered image [8]. Though this process continues, it is thinkable to overwrite
the depiction of binary secret data. Concerning to the grayscale images that pixels have values just single
ranging (0 to 255) with an 8 bits depth, these secret information bits do not convert into binary bits due to their
directly using to substitute the cover image of the objects. With referring to color images which hold 3 routes
red, green, and blue (RGB) besides 24 bits depth, then a cover object (image) can be originally partitioning
into 3 channels just a secret data is embedded early in each channel. Then, these three paths will be merged in
order to create the SI. The LSB bits modification may not be allowed the human visual system (HVS) for
detecting the stego-image. Owing to the fact of LSB substitution method as discrete kind is employed in the
suggested system, the mathematical expression for this method can be prepared with acceptable details. A
mathematical expression aims for providing deeper perception on the focal thought of the scheme in the next
section. The assorted LSB embedding percentage (EP) contains 6.25%, 12.5%, 18.75% in addition to 25%,
that intends 0.5, 1.0, 1.5, and 2 bpp respectively may be utilized depend on the embedded. By using a simple
instance, an inclusive description of the fundamental concept of the LSB steganography basis can be providing.
If a 12.5% EP example is considered, that is mean 1 bpp in each LSB, and can be extended to another EP. With
such implementation, a stego image imperceptibility is decreased, thus making it simple for the HVS for
noticing the stego-image. At this time, image quality will be compromising for data capacity. When data with
larger amount is covered, an image quality degradation is arising. By using LSB methods, the capacity with
high data can be achieved. Figure 1 shows various stego images of Lena in different embedding percentages
[9]-[11].
Through the random addition of 1 to the gray levels on the CI, the pixels of the image are slightly
modified using the LSB-matching. This is done if there is no correspondence between the secret bit and the
LSB of a given pixel, with the values of the pixels maintained within the range of 0-255. There is no difference
between the process of extraction in LSB and LSB-M, this means to use a shared secret key to obtain a
traversing route, as well as to extract the LSB of each pixel for obtaining real embedded bits. A pair of pixels

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 12, No. 1, February 2022: 694-705
696
(𝑃𝑖, 𝑃𝑖 + 1) is used by least significant bit matching revisited (LSB-MR) [12] as an embedding unit that is
manipulated to (𝑃′
𝑖, 𝑃′
𝑖 + 1) in a way to make criteria is satisfied.
{ 𝐿𝑆𝐵 (𝑃′
𝑖) = S𝑖 𝐿𝑆𝐵 (|
𝑃′
𝑖
2
| + 𝑃′
𝑖+1) = S𝑖+1 }
(1)
Where 𝑃𝑖 and 𝑃𝑖 + 1 denotes the embedding unit while the two secret bits are represented by 𝑆𝑖 and S𝑖+1.
With this correlation, the LSB and LSB-M such as irregular artifacts art does not formed in stego images. More
so, with the use of LSB-MR, the rate at which the pixels are modified in can be minimized in variance method
with LSB and LSB-M. The procedure of extraction involves the generation of a traversing path utilizing a
secret key (SK) as well as a quasi-random-number-generator, and afterwards the extraction of two bits from
each of the units of embedding is performed.
12.5% EP=1.5 bpp 25% EP=2 bpp
50% EP=4 bpp 62.5% EP=5 bpp
Figure 1. The imperceptibility of Lena stego-image (SI) using different EP [12]
In a study carried out by Bhardwaj and Sharmab [13], attempts were made to improve the security
and distribute the message within the entire host image. To achieve this, these researchers investigated the
simulation of images through the use of text, and they used LSB to hide information. According to them, the
aim of their study is to provide three level security in which the secret message is complemented, the
complemented secret message is hidden within a cover image pixel that are selected randomly through the use
of a pseudo random number of generator and inverted bit LSB method. Based on the results of their study, their
proposed approach outperforms the conventional LSB as well as the inverse LSB with lower mean square error
(MSE) and higher PSNR.
Hashim et al. [14] in their study, proposed a secure image steganography created by Huffman coding,
distribution of odd/even, and Henon map. The implementation of the proposed framework has been found to
be less complex in contrast to other current methods. Using the Henon map algorithm, the imperceptibility of
the stego-image is increased by using pixel distribution in order to get a better protection method. Prior to the
process of embedding, the secret message is embedded using Huffman coding. Two main reasons that this
method has always been seen efficient: the first is that during the embedding process it is able to check the
correspondence between secret bits with LSB in order to determine 1 and 0, and the second is the segmentation
of the secret message in order to track and map every bit within the stego image. The performance of their
suggested protocol is better than comparing to [15], [16] in terms of PSNR, based on the findings.
Singh and Datar [17] presented an image steganography created with wavelet transformation and RC4
algorithms. A proposed method with cover image is apportioned into 64 blocks of (8×8) with useful wavelet
transform. The secret message here encrypted earlier by embedding with RC4 algorithm is used for enhancing
the security levels. The Stego image now is difficult for detect by HVS attack. Patel and Cheeran [18] had
employed and investigated the steganography technique and advanced encryption standard (AES) algorithm in
order to produce an assessment and comparison into various images format and allows most appropriate
information with this procedure. LSB substitution algorithms were used for implementing this Steganography

697
method. An investigation and evaluation were completed with different parameters like delay, PSNR, MSE, in
addition to absolute mean square error (AMSE).
A new reversible method of data hiding utilizing pixel-value-difference (PVD) with differential
expansion (DE) were presented by Jana et al. (2016). The secret message initially separated into sub-stream
with n bits size in projected method. PVD can be applied for embedding n-1 bits in addition to embed 1 bit by
applying difference-expansion (DE). Lastly, based on the sharing secret key bit stream, a pair of two-stego
pixels will be distributed among dual image. An extracting technique considers a same embed technique to that
of reverse sensing [19].
The securing color image steganography based on LSB was recommended by Al-Tamimi et al. [20].
They used asymmetric key in the image steganography for LSB scheme which contains 32 integers array.
Inserting of data hiding is random in related to hiding message and pixel selection generator, the applying of
transposition is done for each 24-bits block. A security has been improved with LSB substitution technique.
According to the literature review, the majority of current steganography technologies are incapable of
generating high-quality stego-images, leaving the subject vulnerable to identification by human vision systems.
Consequently, the hidden secret can be easily accessed by the attackers, and therefore, cannot be used as an
authentic information in top-secret security systems.
2.2. Random map function
One of the major reasons' steganography is developed is to afford an environment which is secured,
where data can be transmitted over the network by secret messages in stego-images type. The increasing
concern about the security of data is one of the main motivations of this research. Many studies were executed
in the steganography field with a purpose of developed novel approaches through which by steganography a
message can be secured [21]. With respect to enhance data particularity, various studies deal with random
technique due to their higher efficiency and simple using. Advantages of randomizing algorithms can be
summarized as: i) quick and comfort, or may be both in assorted problems, ii) easy employment, and iii) fast
and likely.
Many authors were found in the literature have leveraged random map function compensations, with
each bosses its limitations and strengths. Depending on behavior, various kinds of random maps are existing,
such as, nubasi [22], Arnold scrambles [23], in addition to knight tour [24]. Concerning normal random maps,
the selection numbers can be completed by one parameter at original conditions [25]. Two random maps can
be utilized for the pixels allocations with the aiming of preserving the proposed method. Also, an overlap
between these two ways that guarantee for data which is introduced can be entirely concealed and discovered
the pixels’ path is so hard or almost impossible [23], [24], so that, our method security is guaranteed.
2.3. Huffman coding
The main objective of the Huffman coding algorithm is to reduce the size of text before embedding to
the image. As shown in Figure 2 the Huffman algorithm depends on reducing the frequent letters and gives
them priority code or short path in the Huffman tree. The capacity is an important concept in steganography
method to make the method more robust, such as better method that can hold high amounts of data inside
hosting image while maintaining the quality of the image represented by PSNR. For the concept of data
management and transfer protocol compression of data transferred from sender to receiver is very useful beside
other techniques used in this research. Figure 3 shows a simple flowchart of Huffman coding applied in this
research. High frequency in this case will get less path of visiting to reduce going deep every time and let less
frequency in the deep of the tree, then will gain many clocks in digital word as 0,1.
Figure 2. Reduce text frequency in Huffman coding

 ISSN: 2088-8708
698
Figure 3. Flowchart of Huffman coding
3. THE PROPOSED METHOD
This section provides a graphic description of the proposed method which is proposed in this study
alongside its key modules. The invention of the methodology is further clarified by this graphic representation
of the framework, as then readers can get a better image and a greater understanding of our process. The
suggested method-based steganography varies from other steganography approaches in that it can have high
protection while retaining image quality at a low cost and with a fair payload. The presented work is ideal for
secured transmission application of various secret bits, like electronic-patient-records (EPR) transmission to
the health care centers, and private communication that requests privacy. Figure 4 shows a graphic explanation
for the planned outline.
Figure 4. Overall flow of the proposed scheme

699
This method is made up of four major sub-section which consist: i) image preparation, involving two
parts, first part responsible for partitioning of an image and the second part, responsible for pixel selection, ii)
preparation of secret data, involving the compression (Huffman coding) of secret data prior to the embedding
stage, iii) stage three includes the adaptive hiding of hidden data inside cover images using a data embedding
algorithm, which allows the stego-image to be generated, iv) finally, the secret-data is retrieved from the stego-
image that has been distributed at the receiving terminal using the extraction algorithm. The data could also be
used as appropriate. All of four major stages are summarizing description as follows.
3.1. Image preparation
Image preparation consists of two parts, first part responsible for partitioning of an image and the
second part, responsible for pixel selection. These two parts at image preparation stage are executed
simultaneously before introducing or can hiding text into an image. The cover image consists of
512×512 pixels come from standard dataset and the pixels arranged as a matrix of two dimensions (2D),
grouping these pixels into the blocks for easy management. Selecting one block or group of pixels for
embedding, then selecting pixels inside the block that normally occur in this stage. Steganography system
normally used one parameter random function or other algorithm like knight tour for a random process. Two
control parameters used in random process to achieve the objective of security. Two phases of image division
will perform; first, partitioning the cover image to 64 sub-small images called a block. Select the pixels inside
this block this selection is the most important process to keep the stego image same original as possible.
3.2. Huffman coding
The objective of the Huffman coding algorithm is to decrease the size of text before embedding to the
image. The Huffman algorithm depends on reducing the frequent letters and gives them priority code or short
path in the Huffman tree. High frequency in this case will get less path of visiting to reduce going deep every
time and let less frequency in deep of the tree, then will gain many clocks in digital word as 0,1. Simply can
summarize the procedures of compression text in Huffman coding used with proposed system as Algorithm 1.
Algorithm 1. Huffman coding
Input: array f [1…n] of numerical frequencies or probabilities
Output: binary coding tree with n leaves that has minimum expected code length for f.
Huffman (f[1….n])
T=empty binary tree
Q=priority queue of pairs (i, f[i], i=1….n, with f as comparison key
For each k=1…n-1
i=extract Min(Q)
j=extract Min(Q)
f [n+k]=f[i]+f[j]
insert Node (T, n+k) with children I, j
insert Rear (Q, (n+k), f [n+k]))
Return T
3.3. Embedding algorithm (EA)
The EA is the responsibility for hiding the secrete message within a CI. By supporting of the secret-
key, the embedding algorithm can be able for concealing the encrypted message adaptively inside the LSB
layer. In Algorithm 1, the key steps involved in the presented embedding operation are illustrated. Here,
marking every pixel into block map is the most crucial procedure. This procedure is referred to as embedding
block. The mechanism shown in Figures 5 provides a better view of the core concept of the suggested
embedding algorithm.
Algorithm 2. Embedding algorithm
Input: Cover image (𝐼𝐶
), Stego Key (𝐾𝑆
), Secret data (𝑆𝐷).
initialize 𝐼𝐶
=cover image, 𝑆𝐷=secret data , 𝐾𝑆
=stego key
Apply Huffman using Algorithm 1 coding on on 𝐷𝑆 to get the compression bit stream 𝑀𝐶𝐵𝑆
)
Segment the 𝑀𝐶𝐵𝑆
into groups, each with 16 or 32 bits
Select an appropriate cover image 𝐼𝐶
from dataset of cover images (𝐷𝑆𝐶𝐼
)
Generate random number 1 and arrange it according to HMF-a vector
Select one block of (8x8) blocks via HMF-a vector
Generate random number 2 and arrange it according to HMF-b vector
Select the destination pixel via HMF-b vector
Generate EM vector and arrange it related to Odd/Even
Mark each pixel with LSB and 𝑀𝐶𝐵𝑆
group
If Bit matched>Bit mismatched so that Embedded directly to pixel from secret by step 15
Otherwise, an Inverting secret message will be embedded with step 15
Iteration (Loop) I=1:N
Fetch 𝑀𝐶𝐵𝑆
bits (0 , 1)

 ISSN: 2088-8708
700
a. If MCBS
=0 and Pixel is even, Do no change in 1-LSB layers.
b. If MCBS
=0 and pixel is odd, Do Changes in 2-LSB layer via replace 0 to LSB layer
Else if the 2-LSB layer is full, Do Changes in 1-LSB layers.
c. If MCBS
=1 and pixel is odd, Do no change in 1-LSB layers.
d. If MCBS
=1 and pixel is even, Do changes in 2-LSB layer via replace 1 to LSB layer
Else if the 2-LSB layer is full, Do Changes in 1-LSB layer
I=I+1
Iterate procedure 16 till all the secret-bits are hidden, and a stego-image (SI) is acquired
Output: Stego-Image (IS
)
Figure 5. Examples of embedding for the proposed scheme
The second objective which this study seeks to achieve through the proposed scheme, is to hide the
secret data within an LSB-layer. This implies that in the LSB, if pixels end with 1, they are odd value pixels,
and otherwise, they are even valued pixel. In this case, embedding will involve substituting secret bits based
on pixel (the second layer, a secret bit is replaced by odd pixel, while in the first-layer, the even pixel is used
to replace the secret bit).
3.4. Retrieving algorithm
The retrieving of the concealed secret information about the stego-image is carried out utilize the
retrieving algorithm. For the secret data to be retrieved successfully, different parameters are utilized. Some of
these parameters include Huffman coding, Henon map function (HMF), and stego-keys of data embedded
scheme. The security feature of a proposing framework is complemented by these parameters, thereby making
it difficult for attackers extract secret data. The key steps involved in the proposed mechanism of extraction
are presented in Algorithm 3.
Algorithm 3. Retrieving algorithm
Input: stego image (𝐼𝑆
), stego key (𝐾𝑆
)
Begin 𝐼𝑆
=Stego-Image, 𝐾𝑆
=Secret-Key
Apply random number 1 using HMF-a vector
Select one block of 64 blocks HMF-a vector
Apply random number 2 using HMF-b vector
Select the stego pixel from HMF-b vector
Apply EM vector and arrange it based on odd/even
Mark the LSB of all pixels
Loop start from I=1:N
Reverse the step 15 of embedding process from algorithm 1
Repeat step 9 until extract all secret bits from stego image
Decompression from the resulting bits of step 10 using algorithm1
Re-construct the inventive data from the realized bits
Output: secret message (𝑀)

701
4. EXPERIMENTAL RESULTS AND DISCUSSION
For performing the experiments at this paper, a MATLAB tool together with eight standard grayscale
images which are shown in Figure 6 was utilized for images with (512x512) size were obtained. The different
stego-images of our scheme at EP=2 shown in Figure 6. Many parameters were used in the proposed scheme
using such as PSNR, bpp, embedding capacity (EC) been evaluated.
Lena Lighthouse Pepper Baboon
Zelda House Couple Boat
Figure 6. Cover images used in the proposed scheme
4.1. Benchmarking created on EC, PSNR, and bpp
The embedding capacity EC can be considered as the ratio of message bits number to cover pixels
number [26], [27], and be directly associated with the pixels number utilized in the suggested scheme here as
a various numbers of message bits that can embed by one pixel.
𝐸𝐶 =
𝑇ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑚𝑒𝑠𝑠𝑎𝑔𝑒 𝑏𝑖𝑡𝑠
𝑇ℎ𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑜𝑣𝑒𝑟 𝑖𝑚𝑎𝑔𝑒𝑠′𝑠 𝑝𝑖𝑥𝑒𝑙𝑠 (2)
The diverse payload capacities have been used in this calculation, and offered like percentage with the intention
of being in agreement with a most recent study now. Intended for more explanation, some information is given
for an image 512x512:
a) At 6.25%, 16384 bytes equal, which means every two pixels=16 bits, thus 1/16=6.25% when 1 bit of two
pixels is embedded
b) At 12.5%, 32768 bytes equal, which means every two pixels=8 bits, thus 1/8=12.5% when 1 bit of one
pixel is embedded
c) At 18.75%, 49152 bytes equal, which means every two pixels=16 bits, thus 3/16=18.75% when 1.5 bits of
one pixel is embedded
d) At 25%, 65536 bytes equal, which means every pixel=8 bits, thus 2/8=25% when 2s bit of one pixel is
embedded.
The explanation for using these figures in this analysis is that previous research used different
payloads, so a requirement for having uniform tools to get equal results. In Figure 7, the embedding percentages
utilized in the proposed scheme are presented. The evaluation of image quality can be specified by PSNR that
is evaluated after the embedding process to comparison between original and stego images. A hiding data
process can be considered as an unnoticeable to HVS, PSNR calculation result may be equal or may be greater
than 30 dB [14]. PSNR are measured with applying these equations:
𝑃𝑆𝑁𝑅 = 10𝑙𝑜𝑔10 (
255
𝑀𝑆𝑅
) (3)
where mean square error is MSE, and can be calculated with:

 ISSN: 2088-8708
702
𝑀𝑆𝐸 =
1
𝑚𝑛
∑ ∑(𝑥𝑖𝑗 − 𝑦𝑖𝑗)
𝑛
𝑗=1
𝑚
𝑖=1 (4)
Here, m and n are the images’ sizes, while the cover and stego images are represented with x and y respectively.
During the implementation of the proposed scheme two important stages were carried out on this study, namely
the training and testing stages [27]. In conventional processing of images, the imperceptibility of a SI is
measured by PSNR measures [28]. The consistency of the stego-image has been assessed against the existing
carrier image using the PSNR measures described above. The results of our scheme, and others methods of Wu
and Tsai [29], Wu and Hwang [30], Kumar and Chand [31], Sahu et al. [32], Sahu and Swain's [33] and Qui
et al. [34], are presented in Tables 1-5. More so, Aljuaid and Parah [35] give more results and these schemes
and techniques have been compared with term of EC graphically presented in Figure 8.
Figure 7. Different embedding percentages (EP)
Table 1. Results of the projected scheme at 6.25% and 12.5% of EP
Image (512x512)
Proposed Scheme (6.25%) Proposed Scheme (12.5%)
PSNR EC BPP PSNR EC BPP
Lena 72.01 131.072 0.5 66.44 265.144 10
Lighthouse 72.03 131.072 0.5 66.42 265.144 10
Zelda 72.01 131.072 0.5 66.46 265.144 10
Pepper 72.02 131.072 0.5 66.43 265.144 10
Baboon 72.03 131.072 0.5 66.49 265.144 10
Boat 72.01 131.072 0.5 66.48 265.144 10
House 72.02 131.072 0.5 66.44 265.144 10
Couple 72.02 131.072 0.5 66.43 265.144 10
Average 72.02 131.072 0.5 66.44 265.144 10
Table 2. Results of the scheme at 18.75% and 25%
Image (512x512)
Proposed Scheme (18.75%) Proposed Scheme (25%)
Lena 61.05 393.215 1.5 54.80 542.288 2.0
Lighthouse 61.06 393.215 1.5 54.80 542.288 2.0
Zelda 61.08 393.215 1.5 54.82 542.288 2.0
Pepper 61.18 393.215 1.5 54.10 542.288 2.0
Baboon 61.04 393.215 1.5 54.82 542.288 2.0
Boat 61.05 393.215 1.5 54.80 542.288 2.0
House 61.05 393.215 1.5 54.81 542.288 2.0
Couple 61.09 393.215 1.5 54.82 542.288 2.0
Average 61.07 393.215 1.5 54.85 542.288 2.0

703
Table 3 Results for [30] and [29] techniques
Image (512x512)
Wu [30] (12.5%) Wu and Tsai [28]
Lena 51.05 256.144 1.0 41.45 400.104 1.53
Lighthouse 51.06 256.144 1.0 40.11 411.903 1.57
Zelda 51.08 256.144 1.0 40.01 399.029 1.52
Pepper 51.18 256.144 1.0 40.11 399.140 1.53
Baboon 51.04 256.144 1.0 40.02 419.209 1.60
Boat 51.05 256.144 1.0 39.03 411.674 1.57
House 51.05 256.144 1.0 39.23 401.249 1.53
Couple 51.09 256.144 1.0 39.49 400.670 1.53
Average 51.07 256.144 1.0 39.78 405.238 1.54
Table 4. Results for [31] and [32] techniques
Image (512x512)
Kumar and Chand’s (12.5%) [30] Sahu (18.75%) [32]
Lena 51.27 256.144 1.0 47.83 393.126 1.5
Pepper 51.28 256.144 1.0 47.93 393.126 1.5
Baboon 51.27 256.144 1.0 47.73 393.126 1.5
Boat 51.27 256.144 1.0 47.23 393.126 1.5
House 51.28 256.144 1.0 47.73 393.126 1.5
Average 51.27 256.144 1.0 47.43 393.126 1.5
Table 5. Results for [32] and [33] techniques
Image (512x512)
Sahu (18.75%) [32] Sahu and Swain’s (25%) [32]
Lena 47.32 542.288 2.0 44.08 542.288 2.0
Pepper 47.33 542.288 2.0 44.07 542.288 2.0
Baboon 47.32 542.288 2.0 44.08 542.288 2.0
Boat 47.18 542.288 2.0 44.07 542.288 2.0
House 47.30 542.288 2.0 44.08 542.288 2.0
Average 47.28 542.288 2.0 44.08 542.288 2.0
Lena Lighthouse Pepper Baboon
Zelda House Couple Boat
Figure 8. Different stego-images with EP=5%
The suggested methodology has PSNR for embedding percentage as: EP=6.25% is 72.80 dB, for
EP=12.5% it is 66.60 dB, for EP=18.75 it is 61.20 dB, and for EP=25% it is 54.95 dB. The EC of the proposed
scheme is 131, 072, 265, 144, 393, 216, and 524, 288 bits for EP=6.25%, 12.5%, 18.75 and 25% respectively.
In the studies carried out by Wu and Hwang [30], and Kumar and Chand [31], PSNR of their techniques were
51.08 dB and 51.27 dB, respectively. As seen from the table of comparison, the proposed technique
outperforms that of Wu and Kumar and Chand in terms of PSNR when EP=12.5% with equal EC. The scores

 ISSN: 2088-8708
704
of the PSNR of the proposed scheme when EP=6.25% and 12.5% is presented in Table 3. Based on the results,
the marked images of the proposed scheme are of better quality, thereby confirming its effectiveness. Likewise,
the PSNR and EC of our scheme for EP=18.75% and 25% is better than that of Sahu et al. [32], Sahu and
Swain's [33] respectively.
5. CONCLUSION
This study offers an improving image steganography technique related to odd/even pixel allocation
and with modified the security of secret-data prior hiding. This developed scheme may be established for
enhancing the security level as well as payload capacity for determination the existing complications. A
technique of Huffman coding is utilized for compressing the secret data preceding the embedding. Modified
secret data before hiding that will support to enhance the security and as will capacity. Different payload
capacity has been used with current study and reflected as a percentage to correspond with the researches in
recent studies. The goal of the scheme is to increase imperceptibility by utilizing PSNR measurement so it can
stand against the attack. A good PSNR value in the results shown and these findings confirmed the eligibility
of the proposed scheme. Pixel allocation with high complexities and bits standing by will certify better security
as well as robust imperceptibility associated for the existing scheme.
REFERENCES
[1] M. Al-Mualla and H. Al-Ahmad, “Information hiding: steganography and watermarking,” Multimedia Communication and Signal
Processing (MCSI), 2009.
[2] D. K. Nayak and C. Bhagvati, “A threshold-LSB based information hiding scheme using digital images,” 2013 4th International
Conference on Computer and Communication Technology (ICCCT), 2013, pp. 269-272, doi: 10.1109/ICCCT.2013.6749639.
[3] W. A. Shukur and K. K. Jabbar, “Information hiding using LSB technique based on developed PSO algorithm,” International
Journal of Electrical and Computer Engineering (IJECE), vol. 8, no. 2, pp. 1156-1168, Apr. 2018, doi: 10.11591/ijece.v8i2.pp1156-
1168.
[4] M. M. Kareem, M. Ismail, M. A. Altahrawi, N. Arsad, M. F. Mansor, and A. H. Ali, “Grid based clustering technique in wireless
sensor network using hierarchical routing protocol,” 2018 IEEE 4th International Symposium on Telecommunication Technologies
(ISTT), 2018, pp. 1-5, doi: 10.1109/ISTT.2018.8701720.
[5] A. H. Ali, M. R. Mokhtar, and L. E. George, “Enhancing the hiding capacity of audio steganography based on block mapping,”
Journal of Theoretical and Applied Information Technology, vol. 95, no. 7, pp. 1441-1448, 2017.
[6] B. Wu, M. P. Chang, B. J. Shastri, P. Y. Ma, and P. R. Prucnal, “Dispersion deployment and compensation for optical steganography
based on noise,” in IEEE Photonics Technology Letters, vol. 28, no. 4, pp. 421-424, Feb. 2016, doi: 10.1109/LPT.2015.2496957.
[7] V. Sedighi, R. Cogranne, and J. Fridrich, “Content-adaptive steganography by minimizing statistical detectability,” in IEEE
Transactions on Information Forensics and Security, vol. 11, no. 2, pp. 221-234, Feb. 2016, doi: 10.1109/TIFS.2015.2486744.
[8] A. Alabaichi, M. A. A. K. Al-Dabbas, and A. Salih, “Image steganography using least significant bit and secret map techniques,”
International Journal of Electrical and Computer Engineering (IJECE), vol. 10, no. 1, pp. 935-946, 2020, doi:
10.11591/ijece.v10i1.pp935-946.
[9] O. Evsutin, A. Kokurina, R. Meshcheryakov, and O. Shumskaya, “The adaptive algorithm of information unmistakable embedding
into digital images based on the discrete Fourier transformation,” Multimedia Tools and Applications, vol. 77, no. 21,
pp. 28567-28599, 2018, doi: 10.1007/s11042-018-6055-9.
[10] A. H. Ali and A. D. Farhood, “Design and performance analysis of the WDM schemes for radio over fiber system with different
fiber propagation losses,” Fibers, vol. 7, no. 3, pp. 19, 2019, doi: 10.3390/fib7030019.
[11] S. Dhargupta, A. Chakraborty, S. K. Ghosal, S. Saha, and R. Sarkar, “Fuzzy edge detection-based steganography using modified
Gaussian distribution,” Multimedia Tools and Applications, vol 78, 2019, doi: 10.1007/s11042-018-7123-x.
[12] J. M. Lopes, P. Marrone, S. L. Pereira, and E.M. Dias, “Health 4.0: Challenges for an Orderly and Inclusive Innovation,” IEEE
Technol. Soc. Mag., vol. 38, no. 3, pp. 17–19, 2019, doi: 10.1109/MTS.2019.2930265
[13] R. Bhardwaja and V. Sharmab, “Image steganography based on complemented message and inverted bit LSB substitution,”
Procedia Computer Science, vol. 93, pp. 832–838, 2016, doi: 10.1016/j.procs.2016.07.245.
[14] M. M. Hashim and M. S. M. Rahim, “Image steganography based on odd/even pixels distribution scheme and two parameters
random function,” Journal of Theoretical & Applied Information Technology, vol. 95, no. 22, pp. 5977-5986, 2017.
[15] A. Farhood, M. K. Naji, S. H. Rhaif, and A. H. Ali, “Design and analysis of dual band integrated hexagonal haped microstrip UWB
antenna,” Indonesian Journal Electrical Engineering and Computer Science (IJEECS), vol. 15, no. 1, pp. 294-299, 2019, doi:
10.11591/ijeecs.v15.i1.pp294-299.
[16] M. M. S. A. Al-Momin, I. A. Abed, and H. A. Leftah, “A new approach for enhancing LSB steganography using bidirectional
coding scheme," International Journal of Electrical and Computer Engineering (IJECE), vol. 9, no. 6, pp. 5286-5294, 2019, doi:
10.11591/ijece.v9i6.pp5286-5294.
[17] S. Singh and A. Datar, “Improved hash-based approach for secure color image steganography using canny edge detection method,”
International Journal of Computer Science and Network Security (IJCSNS), vol. 14, no. 7, pp. 82-89, 2014.
[18] F. R. Patel and A. N. Cheeran, “Performance evaluation of steganography and AES encryption based on different formats of the
image,” International Journal of Advanced Research in Computer and Communications Engineering, vol. 4, no. 5, 2015, doi:
10.17148/ijarcce.2015.45140.
[19] H. M. Noman, A. A. Abdulrazzaq, M. M. Kareem, and A. H. Ali, “Improvement Investigation of the TCP algorithms with avoiding
network congestion based on OPNET,” IOP Conference Series: Materials Science and Engineering, vol. 518, no. 5, 2019,
Art. no. 052025, doi:10.1088/1757-899X/518/5/052025.
[20] P. Thakur, S. Kushwaha, and Y. Rai, “Enhance steganography techniques: A solution for image security,” International Journal of
Computer Applications, vol. 115, no. 3, 2015, doi: 10.5120/20133-2220.

705
[21] A. A. Hussein and A. H. Ali, “Comprehensive investigation of coherent optical OFDM-RoF employing 16QAM external
modulation for long-haul optical communication system,” International Journal Electrical and Computer Engineering (IJECE),
vol. 10, no. 3, pp. 2607-2616, 2020, doi: 10.11591/ijece.v10i3.pp2607-2616.
[22] M. Kamath and R. S. Kunte, “Framework for reversible data hiding using cost-effective encoding system for video steganography,”
International Journal of Electrical and Computer Engineering (IJECE), vol. 10, no. 5, pp. 5487-5496, 2020, doi:
10.11591/ijece.v10i5.pp5487-5496.
[23] S. Roy and A. K. Pal, “A robust reversible image watermarking scheme in DCT domain using Arnold scrambling and histogram
modification,” International Journal of Information and Computer Security, vol. 10, no. 2-3, pp. 216-236, doi:
10.1504/IJICS.2018.091469.
[24] A. Banharnsakun, “Artificial bee colony algorithm for solving the Knight’s tour problem,” In International Conference on
Intelligent Computing and Optimization, vol. 866, pp. 129-138, 2018, doi: 10.1007/978-3-030-00979-3_13.
[25] D. Bhattacharyya, T. Kim, and P. Dutta, “A method of data hiding in audio signal,” Journal of the Chinese Institute of Engineers,
vol. 35, no. 5, pp. 523-528, 2010, doi: 10.1080/02533839.2012.679054.
[26] M. Khari, A. K. Garg, A. H. Gandomi, R. Gupta, R. Patan, B. Balusamy, “Securing data in the internet of things (IoT) using
cryptography and steganography techniques,” in IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 50, no. 1, pp.
73-80, Jan. 2020, doi: 10.1109/TSMC.2019.2903785.
[27] M. J. Mnati, R. F. Chisab, A. M. Al-Rawi, A. H. Ali, and A. V. D. Bossche, “An open-source non-contact thermometer using low-
cost electronic components,” HardwareX, vol. 9, 2021, doi: 10.1016/j.ohx.2021.e00183.
[28] S. A. Lafta, A. H. Ali, M. M. Kareem, Y. A. Hussein, A. H. Ali, “Performance simulation of broadband multimedia wireless
networks simulation based on OPNET,” Indonesian Journal Electrical Engineering and Computer Science (IJEECS), vol. 17,
no. 1, pp. 1-9, 2020, doi: 10.11591/ijeecs.v17.i1.pp1-9.
[29] D. Wu and W. Tsai, "A steganographic method for images by pixel-value differencing,” Pattern Recognition Letters, vol. 24,
no. 9-10, pp. 1613-1626, 2003, doi: 10.1016/S0167-8655(02)00402-6.
[30] N. Wu and M. Hwang, “A novel LSB data hiding scheme with the lowest distortion,” The Imaging Science Journal, vol. 65, no. 6,
pp. 371-378, 2017, doi: 10.1080/13682199.2017.1355089.
[31] R. Kumar and S. Chand, “A reversible data hiding scheme using bit flipping strategy,” Journal of Discrete Mathematical Sciences
and Cryptography, vol. 19, no. 2, pp. 331-345, 2016, doi: 10.1080/09720529.2015.1085743.
[32] A. K. Sahu, G. Swain and E. S. Babu, “Digital image steganography using bit flipping,” Cybernetics and Information Technologies,
vol. 18, no. 1, pp. 69-80, 2018, doi: 10.2478/cait-2018-0006.
[33] A. K. Sahu and G. Swain, “A novel n-Rightmost bit replacement image steganography technique,” 3D Research 10, no. 1, 2019,
Art. no. 2, doi: 10.1007/s13319-018-0211-x.
[34] H. Qiu, M. Qiu, M. Liu and G. Memmi, “Secure health data sharing for medical cyber-physical systems for the healthcare 4.0,” in
IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 9, pp. 2499-2505, Sept. 2020, doi: 10.1109/JBHI.2020.2973467.
[35] H. Aljuaid and S. A. Parah, “Secure patient data transfer using information embedding and hyperchaos,” Sensors, vol. 21, no. 1,
2021, Art. no. 282, doi: 10.3390/s21010282.

Security and imperceptibility improving of image steganography using pixel allocation and random function techniques

More Related Content

What's hot (19)

Similar to Security and imperceptibility improving of image steganography using pixel allocation and random function techniques (20)

More from IJECEIAES (20)

Recently uploaded (20)

Security and imperceptibility improving of image steganography using pixel allocation and random function techniques