The development of today's society presents two obvious characteristics: digitalization and networking. Digitalization refers to the storage form of information, which is characterized by large storage capacity and easy editing and copying; Networking refers to the form of information transmission, which has the advantages of fast speed and wide distribution. In the past 10 years, the use and dissemination of digital media information has exploded. People can get digital information and online services quickly and conveniently through the Internet. But at the same time, piracy has become easier, and the management and protection of digital content has become an urgent problem for the industry.
There are essential differences between digital information and analog information, and the traditional scheme of protecting analog information is no longer effective for digital information. Coupled with some general-purpose processors, such as PC, those hardware-based media protection schemes are easy to be broken. In fact, the commonly used encryption technology can only protect the content of the media during the transmission of information from the sender to the receiver. After receiving the information, all data is transparent to users and is no longer protected. In this case, as a potential solution, digital watermarking has been favored by many scholars.
The basic idea of digital watermarking is to hide some meaningful additional information in the original media data, such as audio, video, images and so on. , which is closely combined with the original data and then transmitted together. At the receiving end, the computer extracts the watermark signal for various purposes. Possible applications include digital signature, digital fingerprinting, broadcast monitoring, content authentication, copy control and secret communication. Digital watermarking is considered as the "last line of defense" against multimedia piracy. Therefore, from the watermarking technology itself, it has broad application prospects and great economic value.
General frame diagram of digital watermarking
Basic framework of digital watermarking
A typical watermarking system consists of an embedder and a detector, as shown in the figure. The embedder (formula (1)) generates a real watermark signal according to the information m to be transmitted, and hides it in the media data X to obtain a watermark signal Y. For the sake of security, the generation of the watermark signal usually depends on the key K.
Y may lose some information when it passes through the transmission network, and it will become Y' when it reaches the detector. This kind of channel is uncontrollable and unknowable to both the embedder and the detector, so it can be called an attack channel. The detector is responsible for extracting information from y', as shown in Equation (2). For the detection of unnecessary host signals, we call it blind watermarking, and vice versa. Because of the need of application, blind watermarking has always been the mainstream of research.
Characteristics of digital watermarking
Although the idea of digital watermarking is simple, in order to achieve the purpose of application, it must meet certain performance indicators, among which the more important features include:
Fidelity: Also called invisibility, it refers to the degree to which watermark embedding causes the quality change of the host signal. Since the host signal is mostly multimedia data for people to watch, the watermark should have high fidelity to increase the security of the watermark itself.
Robustness: refers to the survivability of watermark in the process of editing and processing media data. Various operations on media data will lead to the loss of host signal information, thus destroying the integrity of the watermark, such as compression, filtering, adding noise, cutting, scaling and rotation, as well as some malicious attacks.
● Data payload: refers to the amount of information that the watermark signal can transmit under a certain fidelity. Practical applications require that watermarks can convey multi-bit information.
● Security: In applications, there are always people who want to embed, detect or remove watermarks, but others must be restricted from doing the same. This is the security of watermark. In order to achieve security, important information must be kept secret. For example, watermarks are usually generated with keys.
● False alarm rate: refers to the probability of detecting the watermark incorrectly in the host signal without watermark. Obviously, only when the false alarm rate is low enough can the system be used safely and reliably.
The design of watermark must choose the appropriate technology around the above performance indicators. Some features are incompatible, such as invisibility, robustness and information capacity, which must be weighed.
Watermark can be divided into visible watermark and invisible watermark according to fidelity. As the name implies, visible watermarks can be detected by the naked eye. According to the robustness of watermark, it can be divided into robust watermark and fragile watermark. Robust watermarking can resist a certain degree of signal processing; The characteristic of fragile watermark is that any change to media information will destroy the integrity of watermark and make it undetectable. Therefore, robust watermarking is to ensure the integrity of watermark information, while fragile watermarking is to ensure the integrity of media information, and they have their own uses. There is also a kind of watermark, called semi-fragile watermark, which is robust to some operations, but fragile to the modification of important data features.
Development of Digital Watermarking Technology
In the early days, watermark designers were concerned about how to hide information in digital media without being discovered. Therefore, the watermark information is placed in the lowest bit of binary data, and this scheme is collectively called least significant bit modulation. Obviously, in general signal processing, the information of least significant bit is easily lost, and the robustness of the watermark is poor.
Subsequently, a large number of spatial domain watermarking algorithms appeared. Watermark embedding is no longer to modify a single point in the spatial domain, but to modify the characteristics of a point set or an area, such as mean, variance, parity and so on. Patchwork is a typical example of this method. It randomly selects n pairs of pixels (ai, bi) in the image space, adds D to the brightness of pixel ai, and subtracts D from the brightness of Bi. As a result, the average value of the brightness difference between the two groups of pixels is modified to 2d. The theory of mean and statistical hypothesis testing can judge whether the watermark exists or not. However, patchwork can only embed limited information and is sensitive to geometric transformation. The problem of spatial watermarking algorithm is poor robustness to image processing.
Compared with spatial domain, frequency spectrum is a good signal description method. The low frequency component represents the smooth part of the signal and is the main information; The high-frequency component represents the jitter part of the signal, that is, the edge information. The signal analysis and processing is very intuitive and convenient. Spread spectrum watermarking introduces spread spectrum communication theory, which is a very popular design idea of frequency domain watermarking. It regards digital media as a channel, which usually has a wide bandwidth, while the watermark signal embedded as a transmission signal has a narrow bandwidth. Watermark can be extended to multiple frequency points and then superimposed with media signals. In this way, each frequency component only contains a small energy watermark, which not only ensures invisibility, but also destroys the watermark, so it is necessary to superimpose high-amplitude noise on each frequency. This idea is first applied to DCT domain, and then extended to Fourier transform domain and wavelet domain. In addition, in order to give consideration to the fidelity of the watermark, the human perception model is used to control the energy of the watermark at each frequency point so as not to destroy the signal quality, thus forming an adaptive spread spectrum watermark.
Another important watermarking model is communication that regards watermarking as known sideband information. Sideband information refers to information known to the embedder, including media data. The embedder should make full use of sideband information to improve the probability of correct watermark detection as much as possible. This is of great guiding significance to the design of watermark, indicating that the host signal containing watermark should be selected in the area where watermark can be detected, while ensuring a certain fidelity.
At present, the focus of watermarking research is to explore the maximum amount of information that can be embedded and reliably detected in media signals. It applies the knowledge of communication model and information theory with known sideband information. The research of watermarking algorithm focuses on compression domain, that is, JPEG, MPEG and other compression standards, because compression is a technology that must be adopted in information transmission.
Digital watermarking attack technology
Editing and modifying media data will often lead to the loss of information, and the close combination of watermark and media data will also affect the detection and extraction of watermark. We call these actions attacks. Watermark attack technology can be used to test the performance of watermark, which is an important aspect of watermark technology development. How to improve the robustness of watermark and resist attacks is the most concerned issue for watermark designers.
The first generation of watermark performance evaluation system Stirmark contains a lot of signal and image processing operations, which can be divided into:
● Removing attacks: mainly including A/D, D/A conversion, denoising, filtering, histogram modification, quantization and lossy compression. These operations cause information loss of media data, especially compression, which can eliminate redundancy as much as possible under the premise of ensuring a certain information quality, thus removing the watermark.
● Geometric attacks: mainly include various geometric transformations, such as rotation, translation, scaling transformation, cutting, deleting rows or columns, random geometric transformations, etc. These operations change the spatial or time series arrangement of media data, making the watermark undetectable, so it is also called asynchronous attack.
● * * * Collusion attack: The attacker uses multiple watermarked copies of the same piece of media information and uses statistical methods to construct media data without watermark.
● Repeated embedding attack: An attacker embeds his own copyright information into media data that has been embedded with someone else's watermark, thus causing copyright disputes.
The second generation watermarking attack system was proposed by Voloshynovskiy. Its core idea is to estimate the watermark or original media signal by using reasonable statistical model of media data and maximum posterior probability, so as to eliminate the watermark.
The analysis and research of attack technology promotes the innovation of watermark technology, but it also poses one challenge after another to the watermark itself. At present, there is no algorithm that can resist all attacks, especially geometric attacks, which is recognized as the most difficult problem in academic circles, and there is no mature scheme at present.
Digital watermarking products
In the late 1990s, some watermark products began to appear in the world. The first digital watermarking software for copyright protection of still images was introduced by Digimarc Company of the United States, and then the software was integrated into Adobe's Photoshop and Corel Draw image processing software in the form of plug-ins. LavelIt software introduced by AlpVision can hide some characters in any scanned picture for document protection and tracking. SysCop of MediaSec Company uses watermarking technology to protect multimedia content, and wants to prevent illegal copying, dissemination and editing.
The Technical Organization for Copyright Protection (CPTWG) of the United States has set up a special data hiding group (DHSG) to formulate the technical standards for copyright protection watermarking. They put forward a 5C system to protect DVD copyright. IBM has used digital watermarking in the copyright protection system of digital library. Many internationally renowned business groups, such as South Korea's Samsung and Japan's NEC, have also set up DRM technology development projects. In addition, there are some potential application requirements, such as software search and download statistics, web page security early warning, digital TV program protection and confidential file loss prevention.
Some international standards have combined digital watermarking or reserved space for it. The goal of SDMI is to provide an open framework for music playing, storage and distribution. In SDMI specification, a variety of audio file formats are specified, and copyright protection is realized by combining encryption and digital watermarking technology. In the published JPEG2000 international standard, space is reserved for digital watermarking. The upcoming digital video compression standard MPEG-4(ISO/IEC 14496) provides an interface for intellectual property management and protection, allowing the combination of copyright protection technologies including watermarking.
At home, the government attaches great importance to the development of information security industry. The research of digital watermarking is supported by the National Natural Science Foundation and the "863" project. China Information Hiding Seminar (CIHW) has been successfully held for five times since 1999, which has effectively promoted the research and development of watermarking technology. Last year, the government promulgated the People's Republic of China (PRC) Electronic Signature Law, which provided the necessary legal basis for the application of watermarking technology.
Despite the rapid development of digital watermarking technology, it is still far from practical application. Many projects and research are still in the initial and experimental stage, and the existing watermarking products can not fully meet the needs of use. Nowadays, watermarking technology is developing in depth, and some basic technical and legal problems are being solved one by one. It is believed that in the near future, the combination of watermarking and other DRM technologies will completely solve the problem of digital content management and protection.
Small data 2
Application case of multimedia digital copyright protection
DMD is a digital property protection scheme introduced by SafeNet, an American company with a history of 20 years in the security field. It is a DRM product with encryption technology. Chen Hong, vice president of SafeNet Asia Pacific, introduced several successful application cases at the request of reporters.
DMD is mainly used for music and ringtone downloading, video on demand, multimedia content publishing service and recent mobile TV. Basically, customer selection is based on the following points: based on carrier-class performance, it can handle thousands of users at the same time; Can support a variety of DRM technologies at the same time, and has strong support ability for future DRM technologies; High interoperability ensures safe and stable communication between the server and the client; Advanced authorization ability, such as effectively controlling the use of authorization; High integration ensures that DRM platform is not independent, and can be easily integrated into server platform and combined with charging system.
In music download, NPO is SafeNet's customer in France. They are mainly responsible for releasing CD music and putting the released music on the network of FN@C for people to pay for downloading. After the music content is DRM-processed by NPO, it will be released by FN@C (an open portal) and provided for people to pay for downloading. When the end user pays the money, FN@C will encrypt a part of the proof data and submit it to NPO, and NPO will generate authorization for this user.
On the application level of VOD, Germany Arcor is an ISP supplier. With DRM solution, Arcor encrypts audio and video content and provides users with paid services through the Internet and cable TV. The customer paid the money, and Arcor was officially authorized by SafeNet's DRM solution to let the customer enjoy audio-visual services.
In the application of 3G, British BT Lift Company also adopted the DMD scheme of SafeNet. BT LifeTime buys cable TV content (such as sports programs or music programs) from content providers, converts these contents into dab format, encrypts them by DMD and puts them on its platform, and then resells them to wireless operators, providing users of mobile devices with direct payment for downloading content. Provide authorization to customers who pay through SafeNet DMD.
(Computer World 2005 165438+44No. B6 and B7 10/4)
Design and Implementation of MPEG-4 Video Digital Watermarking Technology
Wuhan University Signal and Information Processing Laboratory
With the rapid development of information technology and computer network, people can not only obtain multimedia information conveniently and quickly through the Internet and CD-ROM, but also obtain the same copy as the original data. The resulting piracy and copyright disputes have become increasingly serious social problems. Therefore, the watermarking technology of digital multimedia products has become one of the research hotspots in recent years.
Although digital watermarking technology has made great progress in recent years, its direction is mainly focused on still images. The development of video watermarking technology lags behind that of still image watermarking technology, because a more accurate human visual model including temporal masking effect has not been fully established. On the other hand, due to the emergence of special attacks on video watermarking, some unique requirements different from still image watermarking are put forward.
This paper analyzes the characteristics of MPEG-4 video structure, and gives an application example of an improved video digital watermarking scheme based on spread spectrum.
Introduction of 1 Video Digital Watermarking Technology
1. 1 Introduction of Digital Watermarking Technology
Digital watermarking technology directly embeds some symbolic information into multimedia content through a certain algorithm, but it does not affect the value and use of the original content, and the human perception system cannot perceive or notice it. Different from the traditional encryption technology, digital watermarking technology can not prevent piracy, but can judge whether the object is protected, monitor the spread of protected data, identify the authenticity, solve copyright disputes and provide authentication evidence for the court. In order to make it more difficult for attackers to remove the watermark, most watermarking schemes are enhanced by the encryption system in cryptography. When embedding and extracting the watermark, one key or even several keys are used together. The general method of watermark embedding and extraction is shown in figure 1.
Several aspects to be considered in the design of 1.2 video digital watermarking.
Watermark capacity: The embedded watermark information must be enough to identify the buyer or owner of multimedia content.
Imperceptibility: Digital watermark embedded in video data should be invisible or imperceptible.
Robustness: It is difficult to remove the watermark without obviously reducing the video quality.
Blind detection: Watermark detection does not need the original video, because it is almost impossible to save all the original videos.
Tampering tip: When the multimedia content changes, the watermark extraction algorithm can sensitively detect whether the original data has been tampered with.
Scheme selection of 1.3 video digital watermarking
By analyzing the existing digital video codec system, the current embedding and extracting schemes of MPEG-4 video watermark can be divided into the following categories, as shown in Figure 2.
(1) Video Watermark Embedding Scheme 1: Watermark is directly embedded into the original video stream. The advantage of this scheme is that there are many methods of watermark embedding, and in principle, digital image watermarking schemes can be applied here. Disadvantages are:
Will increase the data bit rate of the video stream;
Watermark will be lost after MPEG-4 lossy compression;
Will reduce the video quality;
For the compressed video, it needs to be decoded first, then embedded with watermark, and then re-encoded.
(2) Video watermark embedding scheme 2: The watermark is embedded into the DC coefficient of DCT in the coding stage (after quantization and before prediction). The advantages of this scheme are:
Watermark is only embedded in DCT coefficients, which will not increase the data bit rate of video stream;
It is easy to design watermarks that resist various attacks;
It can be modulated according to human visual characteristics through adaptive mechanism, and can obtain better subjective visual quality and stronger anti-attack ability at the same time.
The disadvantage is that the compressed video has a process of partial decoding, embedding and re-encoding.
(3) Video watermark embedding scheme 3: Watermark is directly embedded into MPEG-4 compressed bit stream. The advantages are that complete decoding and re-encoding are not needed, and the influence on the whole video signal is small. Disadvantages are:
The constraint of video system on video compression rate will limit the embedding amount of watermark;
The embedding of watermark may have a bad influence on the motion compensation loop in video decoding system;
The design of this kind of algorithm has certain complexity.
Realization of 2 MPEG-4 Video Watermarking
Based on the above scheme, this paper puts forward an improved scheme of spread spectrum digital watermarking technology for MPEG-4 video coding system on the basis of the second scheme, which embeds the watermark information after spread spectrum modulation into the lowest bit of chroma DCT DC coefficient in video stream IVOP(Intra Video Object Plane). This scheme does not need complete decoding, which greatly reduces the computational complexity.
Complexity, improve real-time. At the same time, because the watermark is embedded in the DC coefficient, the watermark has strong robustness on the premise of ensuring the video effect is not distorted.
2. Characteristics of1MPEG-4 video MPEG-4 video codec is based on VOP (Video Object Plane).
In time, VOP can be divided into internal VOP( 1VOP), forward causal prediction VOP(PVOP), two-way non-causal prediction VOP(BVOP) and panoramic ghost VOP(SVOP). IVOP only uses its own information to encode; PVOP uses the past reference VOP for motion compensation predictive coding; BVOP uses past and future reference VOPs for predictive coding of bidirectional motion compensation; SVOP is a static background in a series of moving images. Therefore, the image information of IVOP is relatively independent, which is most suitable for embedding watermark information.
In space, it consists of several macroblocks with the size of 16× 16, and each macroblock contains six subblocks with the size of 8×8. Among them, the basic flow of IVOP coding with four luminance sub-blocks Y, 1 color difference sub-block U and 1 color difference sub-block V is shown in Figure 3.
In order not to be influenced by the quantization process, this scheme embeds the watermark into the quantized DCT coefficients, thus improving the stability of the watermark. In MPEG-4 compression algorithm, the quantization of DCT coefficients is the key, which directly affects video quality and code stream control algorithm. So MPEG-4 provides a standard quantization table for reference. The table is based on the Human Visual Model (HVS). Considering that human eyes are far less sensitive to high-frequency information loss than low-frequency information loss, watermarking is usually embedded in low-frequency information to improve the robustness of watermarking information. In addition, according to the characteristics that human eyes are more sensitive to the change of brightness information than to the chroma information, in order to maintain the video quality to the greatest extent, this scheme embeds the watermark into the chroma (U sub-block) DCT coefficient. Because DCT is the technical basis widely used in multimedia video compression at present, the video watermarking scheme based on DCT has obvious advantages. Embedding watermark information in DCT DC coefficient after IVOP chroma quantization not only does not need to introduce additional transformation to obtain the spectrum distribution of video, but also the watermark information is not affected by DCT coefficient quantization.
2.2 Video digital watermarking algorithm and implementation
In MPEG-4 video, because the DC coefficient of chroma sub-block in IVOP is a robust parameter and always exists in the video stream, this scheme embeds watermark information into the DC coefficient of chroma sub-block DCT of IVOP after being modulated by M sequence (the longest linear feedback shift register sequence). In this way, the watermark information is difficult to remove without affecting the video effect, and the robustness is strong enough. The scheme uses spread spectrum to detect watermarks conveniently and effectively, which can resist all kinds of attacks and interference and has good confidentiality. The key problem is that the DC system of chroma DCT is a very sensitive parameter of visual system. In this scheme, watermarking the DC coefficient of chroma DCT is equivalent to adding a little interference to it. This kind of interference must be kept below a certain threshold, so that the visual system of human eyes can't feel the subtle change of chroma in the video. Experiments show that embedding watermark in the lowest bit of the chroma DCT DC coefficient of IVOP can meet the requirements.
2.2. 1 video digital watermark embedding
The length of pseudo-random spread spectrum sequence is 255(28- 1), and each bit of watermark information is embedded into the lowest bit of DCT DC coefficient (before prediction after quantization) corresponding to IVOP chroma through the modulation of pseudo-random spread spectrum sequence, so it is generally difficult to remove the watermark information without affecting the video effect. At the same time, the error caused by embedding in the lowest bit of DC coefficient is very small.
The pseudo-random spreading sequence generates the following code:
# define M_LEN 255
# Define M Series 8
for(I = 0； I for(I = M _ SERIES； I {
m[I]= m[I- 1]+m[I-5]+m[I-6]+m[I-7]
m[I]= m[I]% 2；
}
The modulation mode of watermark information bit expansion is as follows:
The watermark information bit is 0, and the pseudo-random spread spectrum sequence remains unchanged;
The watermark information bit is 1, and the pseudo-random spread spectrum sequence is inverted.
This process can be realized by XOR operation. The code is as follows:
wmij=wi^m[j]；
/* Each watermark information bit is spread and modulated into 255-bit spread modulation bits */
Here Wi represents the watermark information code stream, and WMij represents the watermark information extended modulation code stream. Let UDCij represent the DC coefficient sequence of video IVOP chroma DCT (after quantization and before DC prediction calculation), and for convenience, one byte represents one bit of binary code stream information.
The watermark embedding process is as follows:
if(WMij)UDCij 1 = 1；
/* Embedding watermark information according to extended modulation code stream */
else UDCij & amp= 0xFFFE
2.2.2 Video Digital Watermark Extraction
Watermark information extraction is the reverse process of watermark information embedding, and the code is as follows:
if(inv _ UDCij & amp； 1)inv _ Wmij = 1；
else inv _ Wmij = 0；
Here inv_UDCij represents the DC coefficient sequence of video IVOP chroma DCT with watermark information (before dequantization and after DC prediction calculation); Inv_WMij represents the detected watermark information extended modulation code stream. When each IVOP chroma sub-block is decoded, 1 bit spread spectrum modulation signal is obtained, and every 255 consecutive spread spectrum modulation signal bits can be demodulated to obtain 1 bit watermark.
Information, the specific analysis is as follows:
XOR the sequence with the same structure and complete synchronization as the original pseudo-random sequence with the obtained 255 consecutive received sequences of spread spectrum modulation signals, and the number of 1 after statistical operation is recorded as OneCount. Since the autocorrelation function of M sequence has only two values (1 and-1/(2n- 1)), it belongs to binary autocorrelation sequence. Therefore, if the data is not attacked and disturbed, OneCount has only two results: 255 or 0. When OneCount=255, the obtained watermark information bits are1; When OneCount=0, the obtained watermark information bit is 0. If the data is attacked or disturbed, OneCount will have many results. According to statistical analysis, when onecount >: 127, the watermark information bit is 1, and (255-OneCount) of these 255 IVOP chroma sub-blocks are attacked or interfered. When a count
3 Analysis of test results
The experimental results show that the longer the length of M sequence, the better the detection effect, but the amount of embeddable watermark information also decreases accordingly. In this scheme, the watermark is only embedded in video IVOP, and PVOP and BVOP are not modified, so it is robust to frame skipping and frame deletion attacks, because IVOP cannot be skipped or deleted. At the same time, because the watermark information is embedded in the DC coefficient of DCT, and the change of DC coefficient will have a great influence on the video effect, the watermark information is embedded in the lowest bit of DC coefficient of DCT in chroma sub-block. This not only greatly reduces the complexity of watermark embedding calculation and saves the time of MPEG-4 encoding and decoding, but also achieves good video effect and invisibility. Statistically speaking, it will not increase the video stream. In addition, watermark extraction does not need the original video. If the watermark information is not attacked, the scheme can accurately extract the complete watermark of the original video; If the watermark information is attacked, according to the nature of spread spectrum demodulation, this scheme can recover the original watermark information to the maximum extent, and count how many IVOP chroma sub-blocks are attacked.
Because DCT is the technical basis adopted by several multimedia video compression standards (H.26 1, H.263, MPEG-4, etc.). Therefore, the watermarking scheme based on DCT has very important research significance and application prospect in video compression. On this basis, this paper proposes a digital watermarking scheme for MPEG-4 video based on spread spectrum. Practice has proved that the scheme can sensitively detect whether the data has been tampered or destroyed without the original video, and has good stability and robustness, thus providing intellectual property protection and preventing illegal acquisition.
This article is taken from the Application of Electronic Technology.
Application: digital watermarking
Message authentication and digital signature can be applied to digital watermarking.
Traditional watermark is used to prove the legality of the contents on paper money or paper, and digital watermark is used to prove the ownership and authenticity of digital products. Digital watermark is digital information embedded in digital products. It can be the author's serial number, company logo, special text, etc.
Digital watermarking is mainly used to prevent illegal copying (indirectly), determine the ownership (author, publisher, distributor and legal end user), determine the authenticity and integrity of the work (whether it is forged or tampered with), confirm the receiver, undeniable transmission, verification of court evidence, forgery identification, identification of file source and version, Web network inspection and monitoring of thieves, etc.
Traditional watermarking is visible to human eyes, while digital watermarking is divided into perceptual and perceptual types.
Perceptible digital watermarking is mainly used to declare the ownership, copyright and source of products on the spot, and plays the role of advertising or restriction. The detectable watermark is usually a light or translucent pattern that is not ugly; For example, when a TV program is playing, the translucent logo of the TV station is inserted in a corner. Another purpose is to distribute works online, such as giving away a low-resolution image with a visible watermark for free. Watermark is often the information of the owner or seller, which provides clues for finding high-resolution original works. If you want to get high-resolution original works, you have to pay. In order to promote their products on the Internet, some companies first distribute reversible visible watermarks, and then use special software to remove the visible watermarks and add invisible watermarks (publishers, distributors, end users and other information). ) when paying for them. It can be seen that watermark has other uses, that is, to save bandwidth, storage space and other reasons, subtitles and subtitles in multiple languages are embedded in VCD, DVD and other movie copies, and the watermark text in each frame is decoded by hardware in real time and displayed on the screen as needed.
The visible watermark in some products reduces the ornamental value of the works more or less, which makes its use relatively limited. The imperceptible watermark has a higher application level and is more difficult to make.
The imperceptible digital watermark, like invisible words in invisible ink technology, is hidden in digital products. The existence of watermark should be based on the principle of not destroying the appreciation value and use value of original data. The digital watermark is embedded in the protected information in some way. When there is a copyright dispute, the digital watermark is extracted through the corresponding algorithm, so as to verify the ownership of copyright. The protected information can be images, sounds, videos or general electronic documents. In order to make it more difficult for attackers to remove watermarks, most watermarking schemes use keys when embedding and extracting watermarks.
Figure 5.7 Watermark Embedding and Extraction
Although digital watermarking technology can't prevent piracy, it can judge whether the object is protected, monitor the dissemination, authenticity identification and illegal copying of protected data, solve copyright disputes and provide evidence for the court.
The design of digital watermarking needs to consider the following aspects:
Robustness: refers to the ability of protected information to resist the loss of hidden information after some changes. Such as channel noise during transmission, filtering operation, resampling, lossy coding and compression, D/ A or A/ D conversion, geometric transformation of images.