| dc.description.abstract | Transmitting data over a public network such as the Internet necessitates
increasing the security of data communications; especially with the highly sensitive
document transfer. Steganography techniques have been introduced and developed to
provide security to these applications. Fundamentally, the steganography goal is not
only to hinder the adversary from decoding a hidden message, but also to prevent an
adversary from suspecting the existence of covert communications. It does not replace
cryptography but rather improves the security using its obscurity features. If one's
suspicious is raised while using a steganography technique, the goal of the latter will be
defeated regardless whether or not a plaintext is revealed. In this research, two
steganography techniques for digital images were developed. The first algorithm
provides a new and efficient steganographic system, called Characteristic Region-Based
Image Steganography (CR-BIS). It combines both the robustness of Speeded-Up Robust
Features technique (SURF) and Discrete Wavelet Transform (DWT) to achieve
characteristic region steganography synchronization. It avoids hiding data in the whole
image by dynamically selecting characteristic regions for the process of embedding.
Such a dynamic manner of region selection increases the security of embedded data.
The experimental results showed that CR-BIS provided stego-images with a good
perceptual quality; this was indicated by the obtained high Peak Signal to Noise Ratio
(PSNR) values, up to 48.30 dB. The second algorithm, namely, an Improved Robust
and Secured Steganography (IRSS), is an improvement of Mali et al.’s algorithm, which
has a reliability defect as some data cannot be retrieved at the extraction phase. IRSS
has overcome the problem of information loss via adopting the concept of the
embedding map. Besides, it has been proved experimentally that IRSS outperformed the
original one in terms of stego-image quality; this was demonstrated by the achieved
PSNR values, which were between (37.28-39.74) dB. In the two proposed algorithms,
the embedded secret information can be correctly recovered without referring to the
original cover-image. In addition, improving the robustness of the stego-system by
Error Correcting Codes (ECC) insertion and adding redundancy bits to the secret
embedded message is defined and evaluated, in order to suggest an appropriate
robustness enhancing method for the proposed algorithms. As a popular ECC, Reed-
Solomon Code (RS-Code) was used to produce correction bits equal to the number of
bits produced by a specific redundancy factor. It has been concluded from the
experimental results that RS-code improved the robustness of CR-BIS algorithm more
than the addition of redundancy does. On the other hand, adding redundancy bits to the
message has a much better effect on the robustness of IRSS algorithm. | en_US |
