Analysis of Digital Watermarking Technology (I)

What is digital watermarking?
Digital Watermark technology is a new direction of research and application in the field of information security technology. It refers to the use of signal processing methods to embed covert marks in digital multimedia data. Such marks are usually invisible and can only be used for special purposes. Detector or reader can extract. The watermark information may be the author's serial number, company logo, special meaning text, etc., and may be used to identify the source, version, original author, issuer, etc. of the file, image or music product, and to monitor the product. The spread of protected data, authentication and illegal copying, resolution of copyright disputes and provision of evidence for the court.
Digital watermarking is closely related to information security, data encryption, etc. Unlike data hiding, digital watermarking requires robustness. Robustness is a very important aspect of watermarking. Different application purposes (such as data tracking, data monitoring and copyright protection, etc.) produce different design purposes and watermarking algorithms. Some watermarking algorithms require raw data to extract embedded watermark information while others do not. In addition, some watermark algorithms that have been published can completely extract the watermark, while others only check whether the watermark exists or not. It should be noted that these two methods are essentially the same. Watermark extraction can be converted to watermark verification and vice versa. Although watermark algorithms are often associated with still images, video, and audio, digital watermarks can also be applied to text, 3D meshes, and face animations.
In the digital watermarking technology, the data volume and robustness of the watermark constitute a basic contradiction. An ideal watermarking algorithm should be able to both hide large amounts of data and resist various signal variations. However, in reality, these two indicators can not always be realized at the same time, and the actual application generally focuses on one of them. If it is for covert communications, the amount of data is obviously the most important. Because the communication method is extremely concealed, the possibility of being tampered with is very small, so the requirement for robustness is not high, but for ensuring digital security, the situation is exactly the opposite. All kinds of confidential data are faced with being stolen and tampered with at any time. Therefore, the requirements for hiding the amount of data are relegated to the inferior position.
The general mode of digital watermarking includes two stages of embedding and detection, extraction, digital watermark generation, embedded algorithm embedding scheme, and the goal is to make the digital watermark find a good compromise between invisibility and robustness. The detection phase is mainly to design a detection algorithm corresponding to the embedded process. The detection result is either the original watermark (such as a character string or icon) or the test result based on statistical principles to determine the presence of the watermark. The goal of the detection scheme is to minimize the probability of misjudgments and missed determinations.
Digital watermark classification <br> According to digital watermark content can be divided into meaningful and non-meaningful watermark. The former means that the watermark itself is also the encoding of a digital image (such as a trademark image) or a digital audio clip; the latter corresponds to only one serial number. The advantage of meaningful watermarking is that if the decoded watermark is damaged due to attack, one can still confirm by visual observation whether there is a watermark. However, if there are several symbol errors in the non-meaningful watermark such as the decoded watermark sequence, only the statistical decision can be used to determine whether the signal contains a watermark.
According to the use of watermarks, it can be divided into ticket security, copyright protection, tampering tips, and covert identification watermarks. Anti-counterfeiting watermarks are special and are mainly used for printing and electronic ticket security. In general, makers of counterfeit coins cannot make excessive modifications to the bill image, so signal editing operations such as scale conversion need not be considered. However, people must consider the situation such as the damage of the document, the pattern blur, etc., and taking into account the rapid mapping requirements, the digital watermarking algorithm for ticket anti-counterfeiting cannot be too complicated.
Copyright logo watermark is currently the most studied type of digital watermark. Digital works are both commodities and knowledge works. This duality determines that copyrighted watermarking mainly emphasizes concealment and robustness, and requires relatively small amount of data.