Adaptive watermarking algorithm based on CGH with strong robustness.

An adaptive watermarking algorithm based on computer generated holography with strong robustness is proposed. Firstly, two original watermarks are converted to holograms using the computer generated holography by the conjugate-symmetric extension. And the holograms are encrypted by the pre-process method. Then, the original image is decomposed into two pyramidal levels by Con-tourlet transform, and the intermediate frequency is transformed into eight directional subbands. Finally, encrypted holograms are embedded into the directional subbands adaptively. Experiments demonstrate that the PSNR of watermarked image is 38. 385 7 dB. The extracted watermarks can be well recognized when JPEG compression quality factor is bigger than 30 and cropping part is less than 30%. The proposed method also can well resist attacks such as Gaussian noise, salt & pepper noise, speckle noise, Gaussian filtering, median filtering and average filtering. And the watermarks extracted from print-scan experiment can be well recognized. In the phase of watermark extraction, original image is not needed. The watermarking technology proposed can be widely used in copyright protection of digital and printed images. [ABSTRACT FROM AUTHOR]