Research Of Digital Image Blind Watermarking Algorithm Based On Compressive Sensing

In today’s informational society, people are more prefer to use Internet and computers to transfer and share digital media information. When people enjoying the convenience which informatization brought to us, they also suffering from the security problems, like tampering, pirating and so on. In order to identify the copyright of the digital productions and protect their security in application, the digital watermarking technology is born. This technology was developed from1990s and has been used in many areas such like images, audios and videos. And now, it has become one of the most favorite topics of information security technology.This paper mainly takes the digital image blind watermarking algorithm as the research object, which focus on the combination with the compressive sensing (CS) theory, thus providing new ideas for merging this technology with the digital watermarking technology. The mainly works and research results are listed as follows:Firstly, this dissertation makes a description of the digital image watermarking technology, including research background, development, frame, classification, characteristic and evaluation criterion, and makes experiments and analysis of the common watermarking attacks. Moreover, it summarizes the advantages and properties of the technologies which used in watermarking algorithms, such as compressive sensing theory. These works have provided valuable data for designing and studying the digital image watermarking.Secondly, based on the computational secrecy property of CS theory, a digital image blind watermarking algorithm is proposed. The main idea of this scheme is embedding the watermark into the CS measurements of the wavelet high frequency coefficients, thus to has a better volume and transparency. The algorithm reaches good security by using the CS theory, random embedding, keys and blind extracting, while meets a balance with the transparency and robustness.Third, based on the recovery property of the CS theory, a digital image dual watermarking algorithm which can be used in copyright protection, content authentication and information recovery is proposed, thus to against the shortcoming of the single watermarking which lack of enough functions. The main idea of this scheme is using the robust watermarking, which based on the singular value decomposition (SVD), as the first level watermark to protect the copyright, while using the zero-watermarking based on CS measurements to authenticate the content and recover the image. The experimental results indicated that the two watermarking of the proposed scheme are noninterference and complement to each other in functions and performance. Robust watermarking ensures the robustness when facing the common attacks, while zero watermarking reduces the degradation bringing from the dual watermark embedding and makes this scheme owning the function of tampering identification and damage recovery. Due to the reason that the zero watermarking can also restore the robust watermarking, so it supports the robust watermarking in a sense. Moreover, the introduction of the third-party, and with the property of blind extraction and computational secrecy or CS, the dual watermarking algorithm researches an outstanding security that can even resist some protocol attacks. Thus, we can get a conclusion that the proposed dual watermarking processes good security and multi-functions.