The Study Of Blind Water Marking Technique Of Digital Images Based On DCT Domain

Blind watermarking technique of digital images is that the watermarking information is secretly hided in a digital image, and then is extracted or checked without the original cover image after the watermarked image is transferred in the Internet. But there are many problems in the current blind watermarking technique, such as the not good robustness and imperceptibility of algorithms, lesser hided watermarking information, and so on. This paper develops its study based on DCT domain surrounding these problems. By virtue of the study, a grayscale meaningful image blind watermarking algorithm and a blind watermarking algorithm based on the property of DCT coefficients are proposed. Then the design principles are simulated and performances are validated through some common attack experiments of images. The former proposes a grayscale meaningful watermarking image embedding algorithm on block discrete cosine transform (DCT) hypo-low frequency AC coefficients of the original image, in virtue of the even-uneven identity of DCT coefficients, JPEG quantifying technique and binary plane theory. Furthermore, PN series is used to ensure watermarking security. The cover image is not needed when extracting the watermarking data. Through using the periodic property of DCT coefficients and the digital image scrambling transformation namely Zigzag Scan, the latter proposes a meaningful binary watermarking image embedding algorithm on block DCT intermediate frequency AC coefficients of the original image. The original cover image is also not needed when extracting the watermarking data. The two algorithms proposed in this paper are more practical than those relying on the comparison between the watermarked and the original image, as well as of good robustness and imperceptibility. All processes, including watermarking embedding, attacking, extracting and verifying, are simulated in MATLAB environment.