An Adaptive Blind Watermarking Algorithm Based on DCT and Modified Watson's Visual Model
Yongzhong Li, Zhu Hua, Ruiqing Yu +2 more
In this paper, an adaptive blind watermarking algorithm based on compressed DCT and Watsonpsilas Visual Mode is proposed. A normalized masking, deduced from Watsonpsilas model according to the principle of maximum possible quantization error, is used to simplify the computation for watermark insertion. Hidden information is adaptively embedded based on the deduced masking and relationship of four-adjacent-blocks. Arnold transformation is used to encrypt the watermark signal with a secret key and the watermark is extracted without the original image. Experimental results indicate that the proposed strategy is resistant to Gaussian noise, mean filter, JPEG compression, etc.