Research On Multimedia Digital Watermark Techniques

The rapid development and wide application of multimedia and network technology have made the storage, processing and distribution of digital multimedia products much easier and faster, and the digital copyright has become more important at the same time. Encryption techniques can be used to protect digital data. However, after the decryption of the data, the data is in the clear and no longer protected. Digital watermarking is another effective solution that is regarded as the last line of defense. This thesis aims at some important problems of watermarking techniques. The main research efforts are list as follow.Firstly some spread spectrum watermark counterfeit algorithms are proposed. In spread spectrum technique, a pseudorandom sequence is used to modulate the watermark signal. The watermark can't be extracted without knowing this pseudorandom sequence. We give some watermark counterfeit algorithms to construct a pseudorandom sequence which can be used to extract one, tens even hundreds bits of counterfeit watermark signals. To avoid those attack, the pseudorandom sequence must be generated using a one-way function, and it is no reliable to get the correlation value using dot product in some case. The watermark counterfeit algorithms are feasible, for many watermark schemes can't ensure that. Experimental results show the validity of those attacks.Secondly a public detect spread spectrum (PDSS) watermarking scheme is given. To insert one bit watermark, a pseudorandom sequence is inserted into the host signal four times, instead of one time in the basic spread spectrum scheme. Based on the precondition that the interference of the host signal is much less than the correlation between the pseudorandom sequences in each signal, as well as the assumption that the pseudorandom sequence and the host signal are statistically independent, the embedded watermark can be extracted. Two kinds of watermark information, private information and public information, are embedded into the host signal using a private key and a public key. The public key can be used to extract the public information. The private key can be used to extract the private information and verify the validation of the public information. The robustness of this scheme is evaluated. To avoid watermark forgery, the robust hash function can be used. Experimental results show the feasibility of our watermark scheme.Thirdly some watermark verification protocols based on bit commitment and zero knowledge proof are proposed. The ownership prover inserts the watermark into the host signal using symmetric watermark technology based on spread spectrum. The watermark detect key is sent to the verifier hiding in bit commitment. By the interactive protocol between the prover and the verifier, the verifier can extract the embedded watermark, but he can't modify, counterfeit or remove it. Protocols are proposed to verify one watermark bit and several...