Securing Visual Data: An R-Channel LSB

for Image Stegnography

Mrs. Shylaja D N, Suvam Sharma, Shreya Kumari, Saurabh Sharma Timilsina, Rishav Upadhyay
Department of Computer Science and Engineering
Bangalore Technological Institute, Bengaluru, India
{shylajadn222@gmail.com, xuvam@hotmail.com, shreya911avm@gmail.com, sharmasaurabh4476@gmail.com,
rishavupadhyay482@gmail.com}

Abstract—In the 21st century, the challenge of securely Image steganography serves as a protective measure for
transmitting information over digital mediums persists due to the sensitive information by concealing it within images, thereby
looming threat of interception by unauthorized entities. evading detection by the human sense of vision. Its
Steganography emerges as a pivotal solution, enabling the applications span diverse fields including communications,
concealment of sensitive data within seemingly innocuous files,
mobile computing, online voting systems, surveillance, and
thereby bolstering security during communication. This research
paper presents an investigation into the efficacy of various least medical applications for securing medical records [9]-[10].
significant bit (LSB) variants within steganographic techniques. A Over the years, numerous image steganography techniques
comparative analysis is conducted across three prevalent image have been developed and implemented to enhance data
formats—PNG, Bitmap, and GIF—assessing performance through security. Among these techniques, the Least Significant Bit
metrics such as mean square error (MSE) and peak signal-to-noise (LSB) method stands out as one of the simplest methods
ratio (PSNR). The results reveal that PNG and Bitmap formats falling under the spatial domain category. This method
exhibit superior performance in concealing data while upholding involves encoding confidential information into the least
image quality compared to GIF. These findings offer valuable significant bits of each pixel within an image. By doing so, the
insights for practitioners and researchers seeking to optimize
LSB technique provides a certain level of confidentiality when
steganographic methods for secure information transmission.
transmitting information over the internet. One of the key
Keywords— Stegobit, Steganography, Stego-image, LSB, PSNR, advantages of the LSB method lies in its ability to achieve low
MSR values of Mean Square Error (MSE) and high values of Peak
Signal-to-Noise Ratio (PSNR). These characteristics make it
I. INTRODUCTION challenging for the human eye to discern any alterations made
to the image, thereby ensuring a level of stealthiness in the
Recent advancements in multimedia technologies have concealment process. Due to these advantages, the LSB
significantly enhanced the efficiency and speed of information method is widely utilized in the development of image
exchange over networks. The global proliferation of the internet, steganography applications, offering a reliable means of
driven by digitalization, has further facilitated data sharing and securing sensitive information during transmission.
utilization through cloud-based infrastructures [1]-[4]. However,
safeguarding information from unauthorized access poses a A K-LSB-based method was implemented for concealing one
formidable challenge in modern societies. To address this image within another image by utilizing the K-least significant
concern, cryptography and steganography techniques have bits in [11]. In [12], An enhanced LSB substitution technique
emerged as pivotal mechanisms for ensuring data security [5]- for image-based steganography was implemented, leveraging a
[6]. local entropy filter for identifying the bit containing the
concealed image. This approach represents a significant
Cryptographic techniques play a crucial role in securing advancement in the field, as it allows for the concealment of
information by encrypting it, thereby rearranging it in a images with minimal distortion and loss of information. The
manner that makes it difficult for intruders to detect and technique employed in this study aimed to bolster security
intercept. Basic authentication methods encompass hashing, measures by incorporating a two-step process. Initially, the
digital signatures, and establishing threshold values for message intended for embedding was encrypted, thereby
embedding secret bits. However, cryptography alone may not adding an extra layer of security. Subsequently, the encrypted
always provide sufficient assurance that confidential message was converted into a stego-image using the
information is entirely protected. Thus, there has been a implemented method. This approach not only enhances data
growing recognition of the need to complement cryptography security but also results in a notable increase in embedding
with steganography [7]-[8]. Indeed, experts commonly capacity while maintaining a high Peak Signal-to-Noise Ratio
advocates for the integration of both steganography and (PSNR).
cryptograhy as multiple layers of security. This multilayered
approach enhances the robustness of data protection In [13], an automated method for securing a secret message was
measures, offering added resilience against unauthorized developed, incorporating a dual level of security. Initially, the
access and inception. secret message underwent encryption using a Java-based
encryption method known as Character Bit Shuffler (CBS),
constituting the first level of security. Subsequently, the
encrypted secret message bits were embedded within an
image utilizing the least significant bit (LSB) technique,
thereby establishing the second level of security. The adopted
method was chosen for its simplicity and its capacity to
preserve the quality of the image. In [14], an enhanced LSB
Steganography approach was introduced, leveraging the
modulus function for embedding data concealment
information. In [15], a method was devised for concealing
secret message bits within the pixels of an RGB true-color
image, utilizing a secret key for enhanced security. Moreover,
the transmission capacity of the image was augmented by
compressing the secret data, originally consisting of seven
bits in ASCII format, into five-bit streams.

This enables the recovery of the original secret messages

embedded by the sender. The depicted processes elucidate the Figure 1: A Typical Cryptographic System
fundamental operations involved in steganography,
facilitating secure communication and data concealment B. The mathematical formulation for the LSB Stenography
between parties. Techniques

The study's findings indicate a notable enhancement in the

image capacity, with an increase of up to 30% achieved through In this study, the LSB substitution method was employed to
the implemented LSB substitution variants. Additionally, the embed text within 8-bit grayscale cover images (GIF) as well as
Peak Signal-to-Noise Ratio (PSNR) demonstrated improvement 24-bit color images (Bitmap and PNG).
by up to 27% compared to existing techniques. Hence, the For an 8-bit grayscale image denoted as Igrayscale of size, PI gray ×
primary objective of this study is to implement various variants QI gray, each pixel can be represented using equation 1 as
of the LSB substitution method, denoted as Stego 1, 2, 3, and 4, follows:
across three distinct image formats: GIF, Bitmap, and PNG.
Subsequently, a comprehensive comparative analysis will be I gray =
conducted to evaluate the performance of these variants.

II. METHODOLOGY
{x ij | 0 ≤ i ≤ PI gray, 0 ≤ i ≤ QI gray, x ij ∈ {0,1,2,…,255} } (1)

A. Overview of the images-based steganography

Fig.1 provides a visual representation of the process involved For a 24-bit color cover image denoted as Icolor of size PIcolor ×
in concealing secret messages within a given medium, along QIcolor is given in equation 2 as follows:
with an overview of the components comprising a typical
steganographic system. The illustration delineates two
Icolor red =
primary processes: embedding and extraction of text-based
messages, carried out at the sender and receiver's ends,
respectively. X red |0 ≤ i ≤ PIcolor red , xred ∈{0,1,2.....,255} (2)
ij
During the embedding process, secret messages are ij |0 ≤ i ≤ QIcolor red ,
concealed within a cover image using a stego-key, resulting
in the generation of a stego-image. Subsequently, the stego- The n-bit of the secret message to be embedded into the 8-bit
image is transmitted over a communication medium to the grayscale and 24-bit color image, denoted as S, is defined by
intended recipient. At the receiver's end, the extraction equation 3.
process involves retrieving the concealed messages from the
received stego-image using the corresponding stego-key. This
enables the recovery of the original secret messages S = {si | 0 ≤ i ≤ n, si ∈{0,1}} (3)
embedded by the sender. The depicted processes elucidate the
fundamental operations involved in steganography,
facilitating secure communication and data concealment
between parties. Furthermore, the secret message is rearranged to form a k-bit
virtual message S ′ , which is defined in equation 4 as follows:
At the receiver's end, the secret message is extracted from the
stego-image using the provided stego-key. Once the

{ { }}
extraction process is completed successfully, the concealed
message is revealed in its original text format. This final step S ' = s ' i|0 ≤ i ≤ n ' , si ∈ 0,1, .. , 2 k −1 (4)
allows the recipient to access and interpret the hidden
information, ensuring the secure and accurate transmission of '
the intended message. Where n = PI gray× QI gray and n = PI color× QI color
Furthermore, a mapping between the original secret message III. RESULTS AND DISCUSSION
(S) and the embedded message (S ′) is established, yielding
equation 5 as follows: A. Computation of peak signal to noise ratio (PSNR)
k=1
Figure 2 illustrates a bar graph representing the computed
s '
i = si ×
∑ k + j×2 k1 j (5) average PSNR for the considered variants (Stego 1, 2, 3, and 4),
j=0 tested with PNG, BMP, and GIF images. The graph demonstrates
that as the number of stego bits used to embed the secret message
Therefore, the embedding process is completed for each pixel, increases, the average PSNR value decreases. This observation
with k (where, k=1,2,3,4) bits of the rearranged message S ′ indicates that as more stego bits are replaced by the message bits,
stored in each pixel. Thus, each pixel stores the k-bit message the stego-image becomes more distorted, posing a potential
for a stego-pixel, which is formulated by equation 6 as follows: security threat to the concealed message.

Xi' = xi mod2k + si (6) Furthermore, the graph reveals that the PSNR for PNG images is
relatively higher for stego 1-3 bits compared to GIF images. The
Lastly, the embedded image is extracted from the PSNR for GIF images is relatively low compared to other image
stego-image using equation 7. formats, suggesting that GIF images may experience more
distortion when used for steganographic embedding. These
findings underscore the importance of selecting appropriate stego
s'= xi mod 2k (7) parameters and image formats to ensure optimal security and
image quality in steganographic applications.

C. Modalities of Data Collection and Criteria for

Evaluation
For the evaluation of the technique, ten images were
obtained for each of the formats, namely GIF, Bitmap,
and PNG, sourced from the internet. These images
were utilized to assess the performance of the embed-
ding process and its impact on image quality. By ana-
lyzing PSNR and MSE values for the stego-images
across the different image formats, the effectiveness of
the embedding technique can be assessed, providing
insights into its performance and suitability for
concealing information within digital images. Figure 2: Bar chart showing the PSNR computation

Mean Square Error (MSE) provides an estimate of the B. Analysis for Mean Square Error (MSE) and Stego Bits
resultant error in the cover image due to the embedded
data. It can be computed mathematically using Equa- The line plot in Fig. 3 depicts the computed average MSE values
tion 8 as follows: across the considered image formats for stego 1-4 bits. It was
observed that the error for the three image formats was very
minimal at stego 1 bit. However, as the number of replacing
N stego bits increases, the errors become more significant. It was
MSE = ∑ (Ci - C ′)2 (8) noted that the errors in GIF images become very significant
i=1
compared to other image formats (PNG and Bitmap) as the
number of stego bits increases. This observation suggests that
Where Ci denotes the size of the stego-image, CiI is the size of GIF images may be more susceptible to distortion when used for
thecover image and N is the size of the original image. PSNR, steganographic embedding, especially with higher levels of data
or Peak Signal-to-Noise Ratio, is indeed utilized to quantify concealment. These findings emphasize the importance of
the degree of distortion introduced by the embedded data. The considering both the stego parameters and the characteristics of
mathematical formula for calculating PSNR is given by the image format when implementing steganographic techniques
equation 9 as follows: to ensure minimal distortion and optimal security.

PSNR(decibel) = 10 log10 ( ) (9)

A high computed value of PSNR indeed indicates that the

image has undergone less distortion due to the embedded data,
suggesting higher image fidelity and better quality.
Conversely, a lower PSNR value suggests that the image has
experienced more distortion, indicating poorer image quality.
Figure 3: MSE computation across all image formats
 CONCLUSION REFERENCES

[1] R. L, Ambika, Birada and V. Burkpalli. Encryption-based

This study has conducted a comparative analysis on the steganography of images by multiobjective whale optimal pixel
performance of the LSB substitution method across three selection. International Journal of Computers and Applications, 1-10.,
different image formats. Throughout the evaluation process, it 2019
was observed that as the number of replacing stego bits
[2] K., Bailey, and K.Curran. An evaluation of image based
increases, there is a consistent relative drop in the PSNR values, steganography methods. Multimedia tools and applications, 30(1),
accompanied by an increase in MSE. These trends indicate a 55-88, 2006
progressive distortion of the images used for hiding the secret
message. [3] R., Bhardwaj, & V. Sharma. Image steganography based on
complemented message and inverted bit LSB substitution. Procedia
Computer Science, 93, 832-838, 2016
It's important to recognize that the security of the embedded
message largely relies on the perceptibility of the image. Since [4] Danny Adiyan, Z., Purboyo, T. W., & Nugrahaeni, R. A. (2018).
steganographic methods are not completely foolproof, it is Implementation of secure steganography on jpeg image using LSB
method. International Journal of Applied Engineering Research,
imperative to exercise caution when concealing sensitive 13(1), 442-448.
information. While steganography can offer a layer of security
for transmitting covert messages, it's essential to assess the [5] X. Li, and J. Wang, A steganographic method based upon JPEG and
particle swarm optimization algorithm. Information Sciences,
trade-offs between security and perceptibility, and to implement
177(15), 3099-3109. 2007
additional encryption or security measures when dealing with
highly sensitive data. Ultimately, a balanced approach to data [6] J. D Teja,. A. C. S., Rao, & S. Dara, A new image steganography
concealment is crucial to ensuring both security and usability in technique for hiding the data in multilayers of the PNG images. Int J
steganographic applications. Ad Hoc Ubiquitous Computer, 2(7), 104-112, 2017.

[7] A. F., Tukiwala, and Degadwala, S. D. (2014). Data Hiding in Image

Based on the results obtained from the study, it is recommended using Multilevel 2-D DWT and ASCII Conversion and Cyclic
that the cover picture used for steganographic embedding should Mathematical Function based Cryptography. International Journal of
have sufficient capacity to accommodate the secret data. Computer Applications, 105(7).
Specifically, the secret data should occupy no more than 20% of [8] F. A., Rafrastara, R., Prahasiwi, E. H., Rachmawanto, & C. A. Sari,
the cover image on average. This ensures that the embedding Image Steganography using Inverted LSB based on 2 nd, 3 rd and 4
process does not excessively distort the cover image, thereby th LSB pattern. In 2019 International Conference on Information and
Communications Technology (ICOIACT) (pp. 179-184). 2019.
maintaining its perceptual quality and reducing the risk of
detection. [9] M., Jain, and A. Kumar., RGB channel based decision tree grey-alpha
medical image steganography with RSA cryptosystem. International
Additionally, it is advisable to minimize the amount of data to Journal of Machine Learning and Cybernetics, 8(5), 1695-1705. 2017
be hidden within the cover image. By keeping the data payload
[10] S., Rustad, A., Syukur, & P. N. Andono. Inverted LSB image
small, the number of pixels required for embedding can be steganography using adaptive pattern to improve imperceptibility.
minimized, mitigating the potential for noticeable distortion in Journal of King Saud University-Computer and Information Sciences,
the stego-image. This approach helps to strike a balance 2019
between embedding capacity and image quality, ensuring that
[11] O., Elharrouss, N., Almaadeed, & S. Al-Maadeed,. An image
the concealed information remains covert while preserving the steganography approach based on k-least significant bits (k-LSB). In
integrity of the cover image. 2020 IEEE International Conference on Informatics, IoT, and
Enabling Technologies (ICIoT) (pp. 131-135), 2021
[12] A., Arora, M. P., Singh, P., Thakral, & N. Jarwal, Image
steganography using enhanced LSB substitution technique. In 2016
Fourth International Conference on Parallel, Distributed and Grid
Computing (PDGC) (pp. 386-389). 2016.

[13] A., AlWatyan, W., Mater, O., Almutairi, M., Almutairi, & A. Al-
Noori, Security approach for LSB steganography based FPGA
implementation. In 2017 7th International Conference on Modeling,
Simulation, and Applied Optimization (ICMSAO), pp. 1-5, 2017

[14] N., Akhtar, V., Ahamad, & , H. Javed A compressed LSB

steganography method. In 2017 3rd International Conference on
Computational Intelligence & Communication Technology (CICT)
(pp. 1-7)., 2017.

[15] A., AlWatyan, W., Mater, O., Almutairi, M., Almutairi, & A. Al-
Noori. Security approach for LSB steganography based FPGA
implementation. In 2017 7th International Conference on Modeling,
Simulation, and Applied Optimization (ICMSAO) (pp. 1-5), 2017.