DIGITAL WATERMARKING OF IMAGE USING 3-LEVEL DISCRETE WAVELET TRANSFORM (DWT), SVD & MEDIAN FILTER

Abstract

Consequent to the tremendous advancement in the field of Information technology and readily availability of tools, it has become increasingly easy for the unknown users to yield unauthorized copies of multimedia data that are easily accessible on the Internet. To genuinely shield those multimedia data on the internet, innumerable techniques including various encryption techniques, stenography techniques, watermarking techniques and information hiding techniques are available. Digital watermarking is a method in which a portion of digital information is embedded into an image and extracted at later stage for the purpose of ownership verification. Such of the watermarks that are very high on the parameters of integrity and robustness are particularly suitable for copyright protection. It is so because they remain inherent with the image under various stages of manipulations and interferences. Digital data, it must be clearly noted, can be embedded both in spatial and frequency domain. In this paper, a new Singular Value Decomposition (SVD) and Discrete Wavelet Transformation (DWT) based technique is proposed for hiding watermark in full frequency band of images. Median filter is used to further accentuate the result. The quality of the watermarked image and extracted watermark is measured using Peak Signal to Noise Ratio (PSNR) and Normalized Correlation (NC) respectively. It is observed that the quality of the watermarked image is distinctly maintained with the PSNR value of 36dB.Robustness of proposed algorithm is tested for various attacks including salt and pepper noise and Gaussian noise, cropping and JPEG compression, amongst others. Decomposition (SVD) based watermarking algorithms it must be noted, have always fascinated scientists and academia due to its inherent simplicity and some other striking mathematical properties. The subservience of this method has been analyzed and implicitly studied by ascertaining the robustness of this algorithm against administering of various geometric attacks to include rotation, scaling, translation (RST). Experimental results so obtained have been matched and deliberated with existing algorithm that visibly throws encouraging results and inferences.