A NOVEL WATERMARKING SCHEME FOR COPYRIGHT PROTECTION BASED ON ADAPTIVE JOINT IMAGE FEATURE AND VISUAL SECRET SHARING

摘要

A visually verifiable watermarking scheme for the copyright protection of the digital image is proposed in this paper. In this scheme, the clustering function of the generalized brain-state-in-a-box neural network (gBSB) and the second-order moment of grayscale histogram are employed to extract a joint image feature, which is constructed by the luminance and the texture characteristics of the protected image, and the image feature can be adaptively adjusted by altering parameters such as the probability P and the threshold T. In addition, the visual cryptography technique is performed to generate the authentication information and realizes both the visual verification and the asymmetric public verification at the same time. The experimental results show that the proposed scheme is highly robust and resilient to some malicious signal processing attacks and geometric attacks. It can be totally applied to the copyright protection of the digital multimedia.