08-05-2012, 10:49 AM
Hybrid Domain in LSB Steganography
pxc3873121.pdf (Size: 321.88 KB / Downloads: 42)
INTRODUCTION
Communication of secret information between two entities can
be achieved through steganography. The term being derived
from Greek words Steganos and graphia meaning covered
writing. Steganography is an information hiding technique
where the secret information to be communicated is embedded
into a cover media such as image, video, audio and text files
such that the hacker looking at the stego image cannot even
think of the existence of the secret information and only it can be
retrieved at the destination by an authorized person. As the
images have redundant insignificant data that can be replaced by
the secret data, maintaining the perceptual quality of stego
object, the images are used as carriers of secret information in
modern steganography.
LITERATURE SURVEY
Takimoto et al., [1] proposed arrangement and detection method
of invisible calibration pattern based on human visual perception
characteristics where the calibration pattern is embedded to blue
intensity in an original image by adding high frequency
component. Zhi-min He Ng et al.,[2] describes steganography
detection using localized generalization error model where the
discrete cosine transform CDCT features and the markov
features are used as inputs of neural networks for detection to
provide better performance. Yifeng Sun Fenlin liu[3] proposed
a method in selecting cover for image steganography by
correlation coefficient to improve the security of steganographic
system. The cover data are modeled as Gauss-Markor process
where the correlation coefficient of two arbitrary data elements
is the exponent of correlation parameter is selected to improve
security. Sun et al.,[4] presented a image steganography method
based on subband coefficient adjustment in BDCT domain for
hiding information.
ALGORITHM
The embedding and retrieval of payload using HDLS algorithm
is discussed in this section.
Problem definition: Given cover image and the payload, the
objectives are:
1. To embed the payload into the cover image to derive
stego image.
2. High robustness with reasonable PSNR to be
achieved.
HDLS Embedding Algorithm: The payload is
embedded in the cover image using LSB technique. The cover
image matrix and payload are divided into two cells each and
the algorithm
PERFORMANCE ANALYSIS
The evaluation parameter PSNR is used as performance analysis
factor to verify the image quality between cover image and stego
image. The algorithm is tested for over one hundred different
kinds of cover images and payloads with different sizes and
formats. A cover image and few payloads are shown in Figure 5
for the demonstration purpose. The PSNR values between cover
image and stego image for different transform domains such as
FFT, DCT, DWT, HDLS with DCT, and HDLS with DWT are
compared in the Table 3 and table 4.
in the spatial domain itself. The HDLS with DWT has better
PSNR compared to HDLS with DCT as well as individual
transform domain techniques. The security to the payload is
better in HDLS because it is the combination of spatial and
transform domain steganographic techniques.
CONCLUSIONS
In modern communication using public channel steganographic
method is adopted for maintaining privacy and secrecy of data.
In this paper HDLS algorithm is proposed. The cover image and
payload are divided into two cells each and RGB components of
cover image cell I are transformed using DWT/DCT by
retaining the components of cell II in spatial domain itself. The
four MSB bits of payload cell I and cell II are embedded into the
second and fourth LSB positions of cover image cell I frequency
coefficients and cell 2 spatial pixel values respectively. The cell
I coefficients are transformed back to spatial domain and then
combined with cell II components to derive stego image in
spatial domain.The Proposed model is observed to have better
PSNR compared to the existing transform domain techniques
with improved security.