IMPLEMENTATION OF TRANSFORM BASED TECHNIQUES IN DIGITAL IMAGE WATERMARKING

Abstract

Digital image watermarking is used to resolve the problem of data security and copyright protection. In many applications of digital watermarking, watermarked image of good quality is required. But there is tradeoff between number of embedded watermark images and quality of watermarked image. This aspect is quite important in case of multiple digital image watermarking. In this case multiple images singular value decomposition based watermarking algorithm performs much better than other transform based methods. This dissertation presents a robust multiple digital images watermarking using singular value decomposition (SVD) method. The results are compared with Discrete Cosine Transform (DCI) based multiple images watermarking method. In case of SVD image watermarking method only singular values are being varied either in single or multiple image watermarking. This helps in preserving the quality of watermarked image. To get more robust watermark against geometrical attack and JPEG compression is main issue in most of the watermarking algorithms. Due to the property of SVD. the SVDbased digital image watermarking has replaced DCT and DWT based watermarking. But here combining the property of DCT and SVD, a hybrid algorithm is introduced which gives much better results. DCT of the whole cover image is taken then segmented in to blocks for singular value decomposition. Watermark image is scrambled using Arnold transform and then it is embedded in singular values. This not only improves the robustness but also security of the watermark. To evaluate the algorithm, this hybrid DCT-SVD based watermarking is compared with the pure SVD based algorithm on the basis of PSNR and robustness. Experimental evaluation shows that this algorithm is able to resist a variety of attacks including JPEG compression, different signal processing attacks and geometric attacks.