Dual Scrambled Image Watermarking Algorithm in DWT-SVD Composite Domain

Dual Scrambled Image Watermarking Algorithm in DWT-SVD Composite Domain

Dharm Singh (Namibia University of Science and Technology, Namibia), Madhuri Agarwal (College of Technology and Engineering, India) and Charu Singh (VIT University, India)
DOI: 10.4018/978-1-5225-2342-0.ch002
OnDemand PDF Download:
List Price: $37.50


There is an increased risk of copyright violation of multimedia data due to the enormous growth of computer networks that provides fast and error free transmission of any multimedia information. A copyright identifier that may contain some information about the lawful owner is inserted in the contents of the image, without sacrificing its quality. The security levels are increased by using a key value and scaling factor for the embedding and extraction process. The dual scrambled watermark using Arnold and Scrambling sequence is embedded by modifying the singular values of the scrambled cover image's DWT middle frequency sub-band. The simulation was performed on MATLAB 7.7.0 with standard database gray scale images of size 512x512 and watermark of size 64x64 using hybrid dual scrambled watermark schemes. The performance analysis is done on the basis of the degree of scrambling and JPEG compression attack using various parameters. The proposed method achieves better imperceptibility and security for the copyright protection methods.
Chapter Preview


Digital multimedia technology has gradually developed and increased the ease of access to digital information. As digital multimedia (video, audio and images) become available for retransmission, reproduction, and publishing over the Internet, a real need for protection against unauthorized copy and distribution has increased, thus gives rise to the digital watermark technique for copyright protection of digital contents. A digital watermark is a visible or invisible identification code that may contain some information about the intended recipient, the lawful owner or author of the original data in the form of textual data or image. On the modification of any data content, could lead to absence or degradation of the watermark.

Thus, an important role can be played by the digital watermarking in providing the evidence of copyright infringements and thus making it traceable for the improper use of protected multimedia. Digital watermarking has a wide-span of practical applications such as digital cameras, medical imaging, image databases, video-on-demand systems, broadcast monitoring, transaction tracking, authentication, copy control, and device control among many others.

Generally, watermark techniques can be broadly divided into:

  • Spatial Domain: The watermark data is embedded in the pixel value by doing changes in the pixel value intensity in this approach (Schyndelet al., 1994).

  • Transform Domain: The watermark is added in the transform domain in order to have imperceptibility as well as robustness. In this method, transform coefficients are modified for embedding the watermark (Nikolaidis and Pitas, 2003).



In this chapter proposed watermarking method is based on DWT and SVD techniques.

1. Discrete Wavelet Transform (DWT)

Wavelet Transform uses wavelets of finite energy. Wavelets are used as basal functions for representing signals in the DWT transform. Wavelets can be realized by iteration of filters with rescaling. They are local in both frequency/scale (via dialations) and in time (via translations). The resolution of the signal is determined by the filtering operations, and the upsampling and downsampling (subsampling) operations determine the scale of an image.

  • Haar Wavelet: It is a step function taking values 1 and -1, on [0; 1/2] and [1/2; 1], respectively. These are good for edge detection and reconstructing binary pulses.

The sequence x[n] denotes the signal in the figure, where n is an integer. G0 denote the low pass filter while H0 denotes the high pass filter. At each level, the high pass filter produces detail information, d[n], while the low pass filter associated with scaling function produces coarse approximations, a[n] (Dhubkarya and Sonam, 2009).

Figure 1.

Three level wavelet decomposition tree

After passing the signal through the half band low pass filter, half of the frequencies get removed and the resolution gets halved but leaves the scale unchanged. This signal is then subsampled by 2 which doubles the scale and the final wavelet coefficients can be obtained after the recursive iterations on the subsampled output of the low pass filter. In other words the DWT can be thought of as consisting of two processes. The filtering is associated with retrieving of the wavelet coefficients corresponding to the current scale (i.e. high pass filter) and the scaling corresponds to sub sampling of the remaining data (i.e. low pass filter sub sampling). Generally the lower frequency sub-bands (LLX) contains most of the image energy and the high frequency sub-bands (HHX) have less energy and cannot be perceived by the human eye but is less prone to attacks.

Complete Chapter List

Search this Book: