Professional Documents
Culture Documents
WOAkjpcv - 19 03 24.
WOAkjpcv - 19 03 24.
Optimization
Abstract- The development of image-based apps and the format. The format is converted into a different, comparable
ease with which information might now be accessed thanks to picture for system files, such as images.
the Internet have made security-related issues more significant
and difficult to solve. Certain media, including pictures, carry The science and art of covert communication known as
information while protecting it from illegal access; this is a steganography focuses on transmitting covert messages by
topic of significant and ongoing research. Steganography is a modifying the original digital media, including images used to
method of clandestine communication that may be used to create stego objects. For a given payload, the statistical
conceal sensitive information. It may be utilized to embed or distributions of the digital media source need to be likely in
hide many types of sensitive data, such as text and picture files, order to guarantee the security of the covert transmission.
using a private key. For example, data hiding in picture is a Steganography is one method of picture security that
method that protects photographs by enclosing sensitive data. concentrates on concealing sensitive information. The process
The primary goal of any picture steganographic model is to of hiding a picture or text—referred to as sensitive image data
minimize image distortion while protecting embedded data. —within another picture or text—referred to as the cover—in
The cover document is the file—also known as the secrete file— order to carry out cryptography and boost security. In recent
that holds the confidential data. The embedded file that was research works, A variety of image steganography techniques
acquired is called the stego file. Therefore, the primary have been used in recent studies. While the others are based
objective of our proposal is to use picture data to develop a on the spatial domain, which has a popular and
unique Steganography model. Image steganography is the
straightforward extraction and embedding method, the
technique of gathering photos and selecting color channels with
DCNN. The WOA approach then uses ALWT to carry out the
majority of approaches are accomplished based on the
embedding procedure by choosing the best pixels from the transform domain with high robustness. The image-based
pictures. Ultimately, the stego pictures are recovered using steganography conceals the keys to improve security.
inverse ALWT. Here, the same DCNN is used to choose the
window from the obtained pictures. The experimental study's
findings demonstrate the effectiveness of embedding over
image data.
I. INTRODUCTION
Information secrecy is one of the most crucial aspects of
communication. There are two primary approaches of
obtaining information confidentiality. Steganography and
cryptography. The use of codes to protect the confidentiality
of information during communication is known as The most effective cover method is image-based
cryptography. The history of cryptography is lengthy. There steganography, which is widely used in human memory
are several approaches and its progression available in. Since applications and is effective. Their encrypted data frequency
several cryptanalysis techniques have been created to detect creates a secret bit change that is undetectable to outside
the secret information, it is no longer sufficient for parties. Previous studies have demonstrated that there are
information to be kept hidden due to technological certain difficulties in concealing the data within image files.
advancements. Therefore, it's imperative to keep the Lower visual quality of the image while performing any
information's existence a secret. The process makes use of modifications, lack of visibility of embedded information on
the distributed images, intra frame distortion drift in image
steganography. The process of secretly adding data to a data sequences, security, resilience, and greater payload of
source without affecting its perceived quality is known as concealed data are some of the issues with data hiding
steganography. The Greek words steganos, which literally systems. When executing the picture The aforementioned
means "covered," and graphia, which means "writing," are the issues with steganography are brought up, which lowers the
source of the term steganography, which refers to covered security level and stego file image quality. The majority of the
writing. Steganography's most well-known application is literature now in publication uses text or image files for digital
hiding a record inside another document. In information steganography. In general, there is a need to promote
hiding, actual data is typically not maintained in its original
II. EXISTING RESEARCH WORKS
steganography approaches on different types of image data.
A. Related works
There are several reasons why data is hidden in images,
including the need to successfully handle image disorientation, In 2023, “Residual Steganography: Embedding Secret
additional room for concealment, and the inability of the Data in Images using Residual Networks” Traditional image
human visual system to discern the concealed data. The steganography techniques typically involve embedding
emergence of deep learning in image vision tasks has led to information directly into the image data, which can often lead
the suggestion of the steganography approach as a means of to noticeable artifacts or degradation of the image quality[1].
optimizing performance rates with respect to accuracy and In 2020, “High-Capacity Information Hiding Based on
efficiency. It uses deep learning-based techniques to improve Residual Network” Image steganography can be divided into
capacity, security, and autonomous generation. Steganalysis is two species: transform domain and spatial domain information
divided into two phases in these respects, such as feature hiding. One is transform domain information hiding, such as
extraction by hand and classification. Support vector machines Discrete Fourier Transform (DFT), Discrete Cosine Transform
(SVM), ensemble classifiers, convolutional neural networks (DCT) The other is spatial domain information hiding
(CNN), recurrent neural networks (RNN), etc. are a few methods, such as Least Significant Bit(LSB)and neural
examples of deep learning approaches. Deep learning network is applied to information hiding, Residual block is
techniques have been applied to image steganography in used to the hidden network to make the network easier to
recent years to raise the communication security level. When optimize and avoid the degradation problem caused by
deep learning is applied to steganalysis, classification and expanding the depth of the network[2].
feature extraction are integrated into a single framework, and
parameter optimization is done to lower complexity and In 2020, “An improved image steganography technique
dimensionality. While many studies concentrate on applying a using edge based data hiding in DCT domain” classified into
novel steganography model for individual data, relatively few two domains i.e., spatial domain and frequency domain, in
studies address the construction of a steganography model. which data is embedded in the edge pixels of the carrier image
Thus, this research proposes a new steganography model using and in the existing image steganography using edge based data
wavelet transform model aided by deep learning and hiding in DCT domain algorithm [3].
optimization method on image data. In 2020, “Steganography Techniques Using Convolutional
The following contributions are made to the Neural Networks” The encoder neural network determines
implementation of a new steganography model. where and how to place the message, dispersing it throughout
the bits of cover image. The decoder network on the receiving
To suggest a new deep learning-assisted ALWT side, which is simultaneously trained with the encoder, reveals
and picture steganography model in order to the secret image [4].
enhance the caliber of stego items.
In 2020, “Image Steganography Analysis Based on Deep
To obtain a picture using the DCNN technique, Learning” The steganalysis approach utilizes the deep features
select the color channel from the photos, and extracted from the stego image and also based on low
improve the stegos' quality. embedding rate images and multi-class steganography [5].
To suggest an ALWT that incorporates the WOA In 2019, Junhong Zhang et al [6] proposed a distortion
algorithm for the purpose of optimizing the assessment method that takes into account more than only
wavelet type in LWT that aids in the embedding inequity effects caused by flipping pixels. However, it relied
of stego files and covers. In the end, inverse on the visual consequences of flipping identical pixels, often
ALWT facilitates the acquisition of the final stego known as joint distortion measurement. Use the syndrome-
pictures. The higher Peak Signal to Noise Ratio trellis code to reduce embedding distortion and embed
(PSNR) rate guarantees the fitness of the messages instead of selecting specific sites for secret
steganography model that was created. messages. The suggested distortion measurement outperforms
the steganography method in terms of performance, statistical
To assess the effectiveness of the developed
security, and image quality.
picture steganography model by comparing it
with multiple methods and using a variety of In 2019, Inas Jawad Kadhim et al [7] investigated the
performance indicators. deployment of digital image steganography. Here, the pivotal
encounter in modeling a steganographic scheme requires
This research work is arranged as follows. The existing
balancing security, resilience, imperceptibility, and bit
research works are evaluated in section II. The deep learning
embedding rate. This paper evaluates traditional image
and wavelet-based image steganography model is illustrated in
steganography methods and introduces fresh contributions
section III. The development of WOA algorithm for image
across several modalities. This paper provides a
embedding selection is shown in section IV. The enhanced
comprehensive overview of image steganography, including
image steganography using ALWT is depicted section V. The
requirements, operations, features, and performance estimates.
results and discussions are explained in section VI. The
Performance evaluation metrics for examining the
conclusion is given in section VII.
steganographic model were also presented. Additionally, we
covered how to select appropriate cover media for specific invisibility, and a high degree of security have all been made
applications and modern steganalysis systems. possible by LWT . However, the noise in the images has an
impact on the performance. With the Optimized Counting-
In 2019, Ahmed A. Abd EL-Latif et al [8] developed a
Based Secret Sharing Scheme, robustness was guaranteed in
method to create substitution boxes (S-boxes) based on
terms of SSIM, MSE, and PSNR, and it requires less
quantum walks. The suggested QWs S-box model's
computing power. However, it has more time complexity
performance was assessed using the S-box evaluation
when figuring out the embedded probability. These difficulties
criterion.
helps researchers to concentrate on putting their new
Furthermore, the results demonstrate that the S-box is steganography model into practice.
suitable for security-related applications. Furthermore, a
unique model for image steganography was developed. To III. DEEP LEARNING AND WAVELET-BASED IMAGE
enhance the security of embedded data, the developed STEGANOGRAPHY MODEL
paradigm combines quantum walks and classical data
concealment. QWs S-box was used to control extraction and A. Proposed image Steganography Model
embedding procedures. The cryptographic QWs S-box ensures
security throughout extraction and embedding stages. The invisible communication science known as
steganography is employed to provide secure communication.
In 2019, Xiaotian Wu and Ching-Nung Yang [9] Another name for it is covered writing, which conceals the
developed SIS approaches using steganography and existence of messages. Digital steganography is used for
authentication for uncompressed photos. They are not suitable secure and confidential communication in the majority of
for domain compression. The model did not relapse a applications worldwide. Various image carriers, such as
deformed stego image into its original model, but the cover pictures, make up this digital media. In order to obtain stego
image was significant. To address these challenges, we objects, this research study investigates different image data.
devised a (k, n) threshold partial reversible Absolute Moment Information about text files or images is concealed within the
Block Truncation Coding (AMBTC) based on the SIS model,
images during the process of image steganography. The game
which includes authentication and steganography. GF (28)
theory model-based image steganography techniques now in
used a polynomial on the basis of the SIS to partition a secret
image into n noise-similar shares. The AMBTC cover picture use have issues with employing more realistic images and
was embedded with parity bits utilizing developed embedding incur significant computational overhead. It allows for the safe
algorithms. Then, n significant stego images were modeled. transmission of sensitive data or the concealment of its
competently handle the shares. Authentication ensured the existence. With the aid of deep learning and wavelet
legitimacy of the Stego picture. Adequate stego photos can techniques, these distinct steganography methods are
drastically alter the secret. integrated to create a new image steganography method.as
shown in Fig. 3.1.
In 2018, Kumar Gaurav and Umesh Ghanekar [10]
developed a steganography algorithm based on the local
reference edge detection model and exclusive disjunction
(XOR) characteristic. Human eyes are less sensitive to
intensity fluctuations in sharp edges than in uniform areas of
an image. The bit plane dependent XOR coding scheme
improves security and capacity in the LSB bits of edge pixels,
resulting in minimal modifications. While the standard edge-
based steganography paradigm is more imperceptible, it has
limited embedding capabilities. Furthermore, the suggested
approach improves embedding capabilities while maintaining
resilience and imperceptibility.
B. Problem statement
The steganography model for images attains the best
results in terms of embedding speed and provides a notable
enhancement in payload and steganography security.
However, it cannot be used to assess joint compression
embedding techniques. Steganography security is provided by
DCNN, and coding efficiency has increased. It has effectively
quantified the altered distortions brought about by intra- and
inter-channels forecasts. However, the costs of inter-block
distortion are higher. Higher processing efficiency, increased
(17)
(20)
Here, the stego and original data are correspondingly
presented by and , the number of Here, the size of the original multimedia data is known as
columns and rows in image data are correspondingly and the size of the data after embedding the
represented as and . Finally, the proposed multi-media information process is considered as .
steganography model is diagrammatically represented in
Fig.5.1. (b) EC: The greatest amount of secret data that can be
incorporated into cover data is its definition. It is dependent
upon the embedding function and the characteristics of the
cover data. It is obtained from Eq. (21) and calculated as a
percentage.
(21)
(9)
(a) (b)
(c) (d)
(e) (f)
Fig. 1. Performance investigation on Image steganography over various optimization algorithms over (a) BR, (b) EC, (c) MSE, (d) NCC, (e) PSNR and (f) SSIM
II. CONCLUSION [9] A. S. Brandao and D. C. Jorge, "Artificial Neural Networks Applied to
Image Steganography," IEEE Latin America Transactions, vol. 14, no.
Using data from image steganography, a new 3, pp. 1361-1366, March 2016.
steganography model has been implemented in this paper. The [10] S. Tan and B. Li, "Targeted Steganalysis of Edge Adaptive Image
DCNN was used to extract the channels from the images in Steganography Based on LSB Matching Revisited Using B-Spline
order to use them for image steganography. Next, the WOA
algorithm was used to choose the best color channels. Fitting," IEEE Signal Processing Letters, vol. 19, no. 6, pp. 336-339,
Ultimately, the information was embedded using ALWT, and June 2012.
the stego image were obtained using inverse ALWT. To [11] Z. Zhang, G. Fu, R. Ni, J. Liu and X. Yang, "A generative method for
achieve higher steganography performance, additional steganography by cover synthesis with auxiliary semantics," Tsinghua
Science and Technology, vol. 25, no. 4, pp. 516-527, Aug. 2020.
improvements to the heuristic techniques will need to be
[12] J. Zhang and D. Zhang, "Detection of LSB Matching Steganography in
made, as the suggested model only observes similar Decompressed Images," IEEE Signal Processing Letters, vol. 17, no. 2,
performance to current ones. pp. 141-144, Feb. 2010.
[13] Ran-Zan Wang and Yeh-Shun Chen, "High-payload image
steganography using two-way block matching," IEEE Signal Processing
References Letters, vol. 13, no. 3, pp. 161-164, March 2006.
[1] Y. Wang, Y. Cao and X. Zhao, "Minimizing Embedding Impact for [14] N. Subramanian, O. Elharrouss, S. Al-Maadeed and A. Bouridane,
H.264 Steganography by Progressive Trellis Coding," IEEE "Image Steganography: A Review of the Recent Advances," IEEE
Transactions on Information Forensics and Security, vol. 16, pp. 333- Access, vol. 9, pp. 23409-23423, 2021.
345, 2021. [15] A. A. Lopez-Hernandez, R. F. Martinez-Gonzalez, J. A. Hernandez-
[2] Yi Chen, Hongxia Wang, Kim-Kwang Raymond Choo, Peisong He, Reyes, L. Palacios-Luengas and R. Vazquez-Medina, "A Steganography
Zoran Salcic, Mohamed Ali Kaafar, Xuyun Zhang, "DDCA: A Method Using Neural Networks," IEEE Latin America Transactions,
Distortion Drift-Based Cost Assignment Method for Adaptive Video vol. 18, no. 03, pp. 495-506, March 2020.
Steganography in the Transform Domain," IEEE Transactions on [16] M. Sharifzadeh, M. Aloraini and D. Schonfeld, "Adaptive Batch Size
Dependable and Secure Computing, 2021. Image Merging Steganography and Quantized Gaussian Image
Emad Elshazly, Abdelwahab Safey, Abouzaid Refaat, O. Zahran, and Steganography," IEEE Transactions on Information Forensics and
Sayed M. Elaraby, "A secure image steganography algorithm based on Security, vol. 15, pp. 867-879, 2020.
least significant bit and integer wavelet transform," Journal of Systems [17] R. J. Mstafa, Y. M. Younis, H. I. Hussein and M. Atto, "A New Video
Engineering and Electronics, vol. 29, issue. 3, pp. 639-649, June 2018. Steganography Scheme Based on Shi-Tomasi Corner Detector," IEEE
[3] Adnan Gutub & Maimoona Al-Ghamdi "Hiding shares by multimedia Access, vol. 8, pp. 161825-161837, 2020.
image steganography for optimized counting-based secret sharing," [18] L. Zhai, L. Wang and Y. Ren, "Universal Detection of Video
Multimedia Tools and Applications, vol. 79, pp. 7951–7985, 2020. Steganography in Multiple Domains Based on the Consistency of
[4] Meenu Suresh & I. Shatheesh Sam "Optimal wavelet transform using Motion Vectors," IEEE Transactions on Information Forensics and
Oppositional Grey Wolf Optimization for video steganography," Security, vol. 15, pp. 1762-1777, 2020.
Multimedia Tools and Applications, vol. 79, pp. 27023–27037, 2020. [19] R. J. Mstafa, K. M. Elleithy and E. Abdelfattah, "A Robust and Secure
[5] K. Niu, J. Li, X. Yang, S. Zhang and B. Wang, "Hybrid Adaptive Video Video Steganography Method in DWT-DCT Domains Based on
Steganography Scheme Under Game Model," IEEE Access, vol. 7, pp. Multiple Object Tracking and ECC," IEEE Access, vol. 5, pp. 5354-
61523-61533, 2019. 5365, 2017.
[6] W. Su, J. Ni, X. Hu and J. Fridrich, "Image Steganography With [20] T. Rabie and M. Baziyad, "The Pixogram: Addressing High Payload
Symmetric Embedding Using Gaussian Markov Random Field Model," Demands for Video Steganography," IEEE Access, vol. 7, pp. 21948-
IEEE Transactions on Circuits and Systems for Video Technology, vol. 21962, 2019.
31, no. 3, pp. 1001-1015, March 2021. [21] S. Liu, B. Liu, Y. Hu and X. Zhao, "Non-Degraded Adaptive HEVC
[7] Meenu Suresh and I.Shatheesh Sam, "Optimized interesting region Steganography by Advanced Motion Vector Prediction," IEEE Signal
identification for video steganography using Fractional Grey Wolf Processing Letters, vol. 28, pp. 1843-1847, 2021.
Optimization along with multi-objective cost function," Journal of King [22] K. Wang, H. Zhao and H. Wang, "Video Steganalysis Against Motion
Saud University - Computer and Information Sciences, August 2020. Vector-Based Steganography by Adding or Subtracting One Motion
[8] X. Zhang, F. Peng and M. Long, "Robust Coverless Image Vector Value," IEEE Transactions on Information Forensics and
Steganography Based on DCT and LDA Topic Classification," IEEE Security, vol. 9, no. 5, pp. 741-751, May 2014.
Transactions on Multimedia, vol. 20, no. 12, pp. 3223-3238, Dec. 2018. [23] K. Tasdemir, F. Kurugollu and S. Sezer, "Spatio-Temporal Rich Model-
Based Video Steganalysis on Cross Sections of Motion Vector Planes,"
IEEE Transactions on Image Processing, vol. 25, no. 7, pp. 3316-3328, of Artificial Intelligence and Information Technology (ICAIIT), pp. 133-
July 2016. 137, 2019.
[24] X. Yi, K. Yang, X. Zhao, Y. Wang and H. Yu, "AHCM: Adaptive [28] M.E.H.Pedersen and A.J.Chipperfield, " Simplifying Particle Swarm
Huffman Code Mapping for Audio Steganography Based on Optimization", Applied Soft Computing, vol. 10, no. 2, pp. 618-628,
Psychoacoustic Model," IEEE Transactions on Information Forensics 2010.
and Security, vol. 14, no. 8, pp. 2217-2231, Aug. 2019. [29] Seyedali Mirjalili and Andrew Lewis, "The Whale Optimization
[25] Abdollahzadeh, B., Soleimanian Gharehchopogh, F., Mirjalili, S. Algorithm", Advances in Engineering Software, vol. 95, pp.51–67,
"Artificial gorilla troops optimizer: A new nature-inspired metaheuristic 2016.
algorithm for global optimization problems," International Journal of [30] Ravipudi Venkata Rao, "Jaya: A simple and new optimization algorithm
Intelligent Systems, vol. 36, issue. 10, pp. 5887–5958, 2021. for solving constrained and unconstrained optimization problems,"
[26] Q. P. Huu, T. H. Dinh, N. N. Tran, T. P. Van and T. T. Minh, "Deep International Journal of Industrial Engineering Computations, vol. 7,
Neural Networks based Invisible Steganography for Audio-into-Image issue. 1, pp. 19-34, 2016.
Algorithm," IEEE 8th Global Conference on Consumer Electronics [31] Nikhitha Keshava Murthy, Shardul Sharma, M. Jasmine Pemeena
(GCCE), pp. 423-427, 2019. Priyadarsini, Ravi Ranjan, Satyaki Sarkar & N. Sardar Basha "Image
[27] M. Anwar, M. Sarosa and E. Rohadi, "Audio Steganography Using Steganography Using Discrete Cosine Transform Algorithm for Medical
Lifting Wavelet Transform and Dynamic Key," International Conference Images," Advances in Automation, Signal Processing, Instrumentation,
and Control, pp. 2349–2358, 2021.