Summary A novel oblivious and highly robust watermarking scheme using Multiple Descriptions (MD) and Quantization Index Modulation (QIM) of the host image is presented in this paper. The watermark is embedded in the Discrete Contourlet Transform domain (CT). Discrete Countourlet Transform (CT) is able to capture the directional edges and contours superior to Discrete Wavelet Transform (DWT). Watermark embedding is done at two stages for achieving robustness to various attacks. This algorithm is highly robust for different attacks on the watermarked image and superior in terms of Peak Signal to Noise Ratio (PSNR) and Normalized Cross correlation (NC). here the part of summary.

http://paper.ijcsns.org/07_book/200802/20080206.pdf

Robust Digital Watermarking Scheme using Contourlet Transform

Editing, reproduction and distribution of the digital multimedia are becoming extremely easier and faster with the existence of the internet and the availability of pervasive and powerful multimedia tools. Digital watermarking has emerged as a possible method to tackle these issues. This paper proposes a scheme using which more data can be inserted into an image in different domains using different techniques. This increases the embedding capacity. Using the proposed scheme 24 binary images can be embedded in the DCT domain and 12 binary images can be embedded in the spatial domain using LSB substitution technique in a single RGB image. The proposed scheme also provides an extra level of security to the watermark image by scrambling the image before embedding it into the host image. Experimental results show that the proposed watermarking method results in almost invisible difference between the watermarked image and the original image and is also robust against various image processing attacks.

/pdf/multiple-binary-images-watermarking-in-spatial-and-frequency-1ly5wp6wyw.pdf

Multiple Binary Images Watermarking in Spatial and Frequency Domains

This paper presents a robust image watermarking scheme for copyright protection of digital images. In this work, Schur decomposition of host image is explored for watermarking . Watermark image is embedded in the two decompositions of Schur decomposition. Schur decomposition is computationally faster and robust to image attacks than Singular Value Decomposition (SVD). Watermark image is embedded in the D component using scalar quantization. A copy of the watermark is embedded in the columns of U matrix using comparison of the coefficients with respect to the watermark image. Since same watermark is embedded in both U matrix and D matrix the robustness is improved. The proposed algorithm is more secure and robust to various attacks, viz., JPEG2000 compression, JPEG compression, rotation, scaling, cropping, row-column blanking, row-column copying, salt and pepper noise, filtering and gamma correction. Superior experimental results are observed with the proposed algorithm over a recent scheme proposed by Chung et al. in terms of Normalized Cross correlation (NCC) and Peak Signal to Noise Ratio (PSNR).

/pdf/on-the-use-of-schur-decomposition-for-copyright-protection-ltg9jt747s.pdf

On the use of Schur Decomposition for Copyright Protection of Digital Images

A novel oblivious and robust image watermarking scheme using Multiple Descriptions Coding (MDC) and Quantization Index Modulation (QIM) is presented in this work. In the proposed scheme, watermark is embedded in two stages to increase the robustness of the watermark against many of the image processing attacks. First stage embedding is based on MDC. Using MDC, the host image is partitioned into two sub images called even description and odd description. Blocked Discrete Hadamard Transform (DHT) is applied for both descriptions. The binary watermark is embedded in the resulting DC coefficients of the odd description using DC coefficients of the even description as the reference. In the second stage, copy of the same watermark is embedded in the first stage watermarked image using QIM. Proposed scheme is characterized with parameters and Genetic Algorithm (GA) is used for parameter optimization and thereby performance improvement is achieved over the existing methods in terms of Peak Signal to Noise Ratio (PSNR) and Normalized Cross correlation (NCC). Experimental results show that, this algorithm is highly robust for many image attacks on the watermarked image.

/pdf/an-oblivious-and-robust-multiple-image-watermarking-scheme-2aicja9d1n.pdf

An Oblivious and Robust Multiple Image Watermarking Scheme Using Genetic Algorithm

In the past half century, the advancement of internet technology has been rapid and widespread. The innovation provides an efficient platform for human communication and other digital applications. Nowadays, everyone can easily access, copy, modify and distribute digital contents for personal or commercial gains. Therefore, a good copyright protection is required to discourage the illicit activities. On way is to watermark the assets by embedding an owner's identity which could later on be used for authentication. Thus far, many watermarking techniques have been proposed which focus on improving three standard measures, visual quality or imperceptibility, robustness and capacity. Although their performances are encouraging, there are still plenty of rooms for improvements. Thus, this study proposes a new watermarking technique using Least Significant Bit (LSB) insertion approach coupled with Hadamard matrix. The technique involves four main stages: Firstly, the cover image is decomposed into three separate channels, Red, Green and Blue. Secondly, the Blue channel is chosen and converted into an eight bit stream. Thirdly, the second least signification bit is selected from the bit stream for embedding. In order to increase the imperceptibility a Hadamard matrix is used to find the best pixels of the cover image for the embedding task. Experimental results on standard dataset have revealed that average PSNR value is greater than 58db, which indicates the watermarked image is visually identical to its original. However, the proposed technique suffers from Gaussian and Poisson noise attacks.

An improvement of RGB color image watermarking technique using ISB stream bit and Hadamard matrix

A modified digital image watermarking scheme based on the combination of spatial and transform domains is proposed. The aim of this work is to achieve robustness for the vital part of the watermark image without scarifying the embedding capacity of the watermark image. The watermark image is split into two parts depending on the vital information. The perceptibility of the watermarked image is improved by hiding the vital part of the watermark image in the blocks of the host image having highest variance in transform domain. The remaining part of the watermark image is watermarked in the spatial domain. The proposed algorithm is more secure for data hiding and robust for various attacks, viz., cropping and noise. The performance of this algorithm is noteworthy in terms of normalized cross correlation (NCC) and peak signal to noise ratio (PSNR). Additional security is achieved by permuting the watermark image. In this work Hadamard transform is used to hide the vital part of the watermark image in the host image.

S. Srinivas Kumar

Papers

Digital image watermarking in dual domains