Robust image-adaptive watermarking using an adjustable dynamic strength factor

This paper presents a novel spread spectrum based image-adaptive watermarking technique where selective embedding of watermark has been done in small building blocks of the image, depending upon the respective entropies of the blocks. Novelty of this method lies in the flexibility of its strength factor which can be adjusted easily in order to obtain a required level of visual quality of the watermarked image in terms of PSNR. Easy adjustability of the strength factor has helped in experimentally demonstrating the inverse relationship between imperceptibility and robustness of the watermark. Positions of embedded blocks, means of low-frequency wavelet coefficients and size of watermark have been used as side information for decoding of the watermark. Applied technique has shown exceptional resistance to some of the common watermarking attacks such as JPEG compression, median filtering, Gaussian filtering and AWGN noise attacks.