An Improved Methodology for Data Hiding In Images Using Haar Transformed, LSB Replacement Method and Modified PVDMF 1

Steganography is a technique of concealing secret text or image within another image.In this paper, abetterand efficient steganography methodfor hiding imagesand/or text into a color image is proposed. The proposed techniqueis based on two irreversible methods of spatial domain Least Significant Bit (LSB) and Pixel Value Differencing and Modulus Function (PVDMF 1) and one Haar Wavelet Transformation of the frequency domain. The proposed method works in four stages: at the first stage, Haar Wavelet Transformation is applied on the input image and transformed Haar coefficients are obtained. At the second stage, Haar coefficients and secret image are converted into binary form, before applying the LSB replacement method on the green channel by replacing the first two Most Significant Bits (MSB) of the input image withthe last two LSB bits of the secret image. Inthe third stage, modified PVDMF 1 is applied on intermediate stego imagesafter hiding the image by dividing the image into two non-overlapping consecutive pixels of pixels for hiding text. Inthe fourth stage, Inverse Haar Wavelet Transformation is applied to get the stego image. The proposed method offers high hiding capabilitywith goodvisual quality. The visual quality of the stego image is measured by using three widely used objective quality assessment matrices PSNR, RMSE,and SSIM. Furthermore, Regular/Singular (RS) analysis is applied on stego imagesto check theirresistanceagainst RS attack. The hiding capacity and quality assessment results are compared with some existing methods. It is evidentfrom the comparison that the proposed processis superior in termsof visual quality and hiding capacity.

Hardware Implementation of a Visual Image Watermarking Scheme Using Qubit/Quantum Computation Through Reversible Methodology

In this chapter, hardware implementation of an LSB replacement-based digital image watermarking algorithm is introduced. The proposed scheme is developed in spatial domain. In this watermarking process, data or watermark is implanted into the cover image pixels through an adaptive last significant bit (LSB) replacement technique. The real-time execution of the watermarking logic is developed here using reversible logic. Utilization of reversible logic reduces the power dissipation by means of no information loss. The lesser power dissipation enables a faster operation as well as holds up Moore's law. The experimental results confirm that the proposed scheme offers high imperceptibility with a justified robustness.

Secure Electronic Voting Using a Hybrid Cryptosystem and Steganography

Conventional voting techniques have been employed over the years in most elections. However, each of these techniques has attendant short comings. The existing conventional voting systems have been subjected to gross abuse and irregularities. Electronic voting (e-voting) which is emerging as an alternative to these conventional voting systems, though highly promising, is not also totally free of flaws; information security issues bordering on privacy, integrity and verifiability of the electronic ballots casted are still significant in most implementations of e-voting systems. In this paper, we developed a security scheme that was based on a hybrid Rivest-Sharma-Adleman (RSA) algorithm and Advance Encryption Standard (AES) algorithm with Least Significant Bit (LSB) replacement algorithm for securing ballot information in an e-voting system. An e-voting system with multiple voting channels which includes poll site voting, mobile voting and remote internet voting was developed in the contextual Nigerian scenario. The electronic ballots casted via the system during experimental usage were encrypted using the hybrid RSA-AES algorithm to circumvent the problems of privacy, integrity and verifiability associated with many e-voting systems. The encrypted ballots were then hidden using LSB replacement algorithm to improve their integrity. The image quality of the cover image used and stego image obtained from the security scheme developed was quantitatively assessed using Peak Signal to Noise Ratio (PSNR), Mean Square Error (MSE) and Correlation. Furthermore, a performance evaluation based on users' perception of the developed e-voting system showed that the system satisfied most of the generic security requirements for electronic voting.

Cross-Validation and Blind Feature Analysis of 25 Percent Embedding on JPEG Image Format using SVM

This paper provides a result assessment of traditional JPEG picture extraction function steganalysis compared to a cross-validation picture. Four distinct algorithms are used as steganographic systems in the spatial and transform domain. They are LSB Matching, LSB Replacement, Pixel Value Differencing and F5.A 25 percentage of embedding with text embedding information is considered in this paper. The characteristics regarded for evaluation are the First Order, Second Order, Extended DCT characteristics, and Markov characteristics. Support Vector Machine is the classifier used here. In statistical recovery, six distinct kernels and four distinct sampling techniques are used for evaluation.

Blind Steganalysis for JPEG Images using SVM and SVM-PSO Classifiers

Blind steganalysis or the universal steganalysis helps to identify hidden information without previous knowledge of the content or the embedding technique. The Support Vector Machine (SVM) and SVM- Particle Swarm Optimization (SVM-PSO) classifiers are adopted for the proposed blind steganalysis. The important features of the JPEG images are extracted using Discrete Cosine Transform (DCT). The kernel functions used for the classifiers in the proposed work are the linear, epanechnikov, multi-quadratic, radial, ANOVA and polynomial. The proposed work uses linear, shuffle, stratified and automatic sampling techniques. The proposed work employs four techniques for image embedding namely, Least Significant Bit (LSB) Matching, LSB replacement, Pixel Value Differencing (PVD) and F5 and applies 25% embedding. The data to the classifier is split as 80:20 for training and testing and 10-fold cross validation is carried out.