A robust data embedding using perceptual adaptive watermark

In this paper, a new watermarking method to embed robust data against JPEG compression and noise without degrading an image quality is proposed. To guarantee that embedded data are invisible, the proposed method adaptively determines a different size of watermark for each frequency coefficient using human visual system models. The position to be watermarked is chosen in low frequencies using a Pn-sequence to prevent the embedded data from being removed or changed. Experimental results show that the proposed method is better than previous works in terms of bit error of a detected data against JPEG compression.