Blind image watermarking based on discrete fractional random transform and subsampling
This paper proposes a blind watermarking scheme based on discrete fractional random transform. The watermark information can be a binary sequence, a gray level image or a set of decimal fractions sampled from a given source signal. The host image is subsampled into four subimages, and the high correlations among their discrete fractional random transform coefficients are exploited for watermark embedding. Based on this self-reference strategy, the watermark can be extracted without the aid of the host image. As a fragile watermarking technique, our scheme can be used in tamper detection. Besides, it can be used in self-embedding for a large payload is provided. Meanwhile, security of the watermark is preserved due to the randomness of the discrete fractional random transform. Experimental results demonstrate the effectiveness of our scheme.
, vol. 122, no. 4, pp. 311-316, 2011