Content adaptive watermarking using a 2-stage predictor

Digital watermarking is one of the solutions to protect intellectual properties and copyright by hiding information, such as a random sequence or a logo, into digital media. In this paper, a new watermarking scheme is proposed. The embedding process takes place in the spatial domain. A bi-level logo is embedded into digital media by comparing the absolute difference between the original pixel value and the predicted pixel value, which is computed and predicted by using the concept of activity measurement, followed by spatial varying filter (SVF). The logo will be extracted from a possibly corrupted image, without the help of original uncorrupted image. The new proposed algorithm can withstand the geometric attacks as well as common signal processing, and has a high payload capacity.

[1]  Ioannis Pitas,et al.  Copyright protection of images using robust digital signatures , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.

[2]  Ingemar J. Cox,et al.  Secure spread spectrum watermarking for multimedia , 1997, IEEE Trans. Image Process..

[3]  Mauro Barni,et al.  DCT-based watermark recovering without resorting to the uncorrupted original image , 1997, Proceedings of International Conference on Image Processing.

[4]  Oscar C. Au.  Fast ad-hoc inverse halftoning using adaptive filtering , 1999, 1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258).

[5]  Walter Bender,et al.  Techniques for data hiding , 1995, Electronic Imaging.

[6]  Ingemar J. Cox,et al.  Secure spread spectrum watermarking for images, audio and video , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.

[7]  Chengyun Yang,et al.  Reversible data hiding based on wavelet spread spectrum , 2004, IEEE 6th Workshop on Multimedia Signal Processing, 2004..