An SVD Based Location Specific Robust Color Image Watermarking Scheme Using RDWT and Arnold Scrambling

Abstract

In recent times, effectiveness of the robust image watermarking scheme almost successfully solves the authenticity and copyright protection problem of images. Among several existing robust watermarking schemes, very few have been designed for color images. In this paper, the authors have proposed a redundant discrete wavelet transform (RDWT) and singular value decomposition based hybrid robust color image watermarking method where scrambled watermark is used for embedding purpose to provide extra security to authentic logo. At first, RGB color image is converted to YCbCr color image and then the Y component of YCbCr color space is used to insert Arnold scrambled grayscale watermark. In the next step, Y component of the YCbCr color model is decomposed into non overlapping blocks and subsequently RDWT are applied to each block. For better imperceptibility and effectiveness, in this proposed algorithm some singular values of RDWT transformed Y component blocks are modified by non overlapping decomposed and scrambled watermark blocks. Robustness of the proposed method is experimented against common geometric transformation attacks (like rotation, flip operation, cropping, scaling, shearing and deletion of lines or column operation etc.), common enhancement technique attacks (like lowpass filtering, histogram equalization, sharpening, gamma correction, noise addition etc.), jpeg compression attacks, some combinational attacks and satisfactory results are achieved. For better evaluation, proposed scheme is verified against standard benchmark software “Checkmark”.