Tongxin xuebao (Feb 2022)
Discriminative zero-watermarking algorithm based on region XOR and ternary quantization
Abstract
Aiming at the problem that the current zero-watermark algorithm is insufficient in distinguishing the zero-watermark generated by the attacked image and the false-alarm image, a high-resolution zero-watermarking algorithm based on regional XOR and ternary quantization was proposed.In the process of zero watermark generation, the spatial and frequency domain features of the image were first extracted, the carrier feature and copyright watermark were redundantly constructed, and the frequency domain feature was used as an index to divide the redundant watermark into two regions, which were respectively related to the spatial domain features and the copyright watermark.The chaotic matrix was XOR, and the two regions after XOR were combined to obtain zero watermark.In the copyright authentication process, the airspace and frequency domain features were extracted from the attacked image, and the frequency domain feature was used as an index to divide the zero watermark into two regions, which were XOR with the airspace feature and the chaotic matrix respectively, and the redundantly constructed watermark was extracted.The image, after de-redundancy operation, was quantized into a matrix with only three values as the result of copyright authentication.The experimental results show that the proposed algorithm can effectively solve the false alarm problem of image zero watermark, and at the same time has high robustness under various attacks, the feature extraction algorithm can be arbitrarily replaced, and has high scalability.
