Problem description:
How to encrypt files with digital seed watermarking technology?
Analysis:
Digital watermarking technology directly embeds some tag information related or unrelated to multimedia content, but it does not affect the use value of the original content and is not easy to be perceived or noticed by human perception system. Through these hidden information in multimedia content, we can confirm the creator, buyer or authenticity of the content. Digital watermarking is an important research direction of information hiding technology.
As a digital watermarking technology, it mainly has the following characteristics:
-Security: The information of digital watermark should be secure and difficult to tamper with or forge. At the same time, there should be a low false detection rate. When the original content changes, the digital watermark will also change, so as to detect the change of the original data; Of course, digital watermarking also has strong anti-duplication ability.
-Concealment: Digital watermarking should be imperceptible and should not affect the normal use of protected data; Will not be downgraded;
-Robustness: It means that the digital watermark can still maintain partial integrity and be accurately identified after going through various unintentional or intentional signal processing processes. Possible signal processing processes include channel noise, filtering, digital-to-analog and analog-to-digital conversion, resampling, clipping, displacement, scale change and lossy compression coding. Fragile watermark, a digital watermark mainly used for copyright protection, is mainly used for integrity protection. This watermark also embeds invisible information in the content data. When the content changes, the watermark information will also change, so that we can identify whether the original data has been tampered with.
-Watermark capacity: The embedded watermark information must be enough to show the identification information of the creator or owner of multimedia content or the serial number of the buyer, which is conducive to solving copyright disputes and protecting the interests of the legitimate owners of digital property rights. Especially in the field of covert communication, there is a great demand for watermark capacity.
Classification of digital watermarking
-1. Divided by characteristics
According to the characteristics of watermarking, digital watermarking can be divided into robust digital watermarking and fragile digital watermarking. Robust digital watermarking is mainly used to identify copyright information in digital works. By using this watermarking technology, the identification information of the creator and owner or the identification of the buyer (that is, the serial number) is embedded in the data of multimedia content. In the event of copyright disputes, the information of the creator or owner is used to identify the copyright owner of the data, while the serial number is used to track the users who violate the agreement and provide multimedia data for piracy. Digital watermarking for copyright protection requires strong robustness and security. Besides surviving in general image processing (such as filtering, adding noise, replacing and compressing), it also needs to be able to resist some malicious attacks.
Fragile watermarking, contrary to the requirement of robust watermarking, is mainly used for integrity protection. This watermark also embeds invisible information in the content data. When the content changes, the watermark information will also change, so that we can identify whether the original data has been tampered with. Vulnerable watermark should have strong immunity (robustness) to general image processing (such as filtering, noise adding, replacement, compression, etc.). ), at the same time, it must have strong sensitivity, that is, it must allow a certain degree of distortion and be able to detect distortion. It must be sensitive to the change of signal, and people can judge whether the data has been tampered with according to the state of fragile watermark.
-2. According to the media attached to the watermark.
According to the media attached to the watermark, we can divide the digital watermark into image watermark, audio watermark, video watermark, text watermark and mesh watermark of 3D mesh model. With the development of digital technology, more kinds of digital media will appear, and the corresponding watermarking technology will also be produced.
-3. According to the detection process.
According to the detection process of watermark, digital watermark can be divided into plaintext watermark and blind watermark. The plaintext watermark needs the original data in the detection process, while the blind watermark detection only needs the key, not the original data. Generally speaking, plaintext watermarking is robust, but its application is limited by storage cost. At present, most of the digital watermarks studied by academic circles are blind watermarks.
-4. Divide by content
According to the content of digital watermarking, watermarking can be divided into meaningful watermarking and meaningless watermarking. Meaningful watermark means that the watermark itself is also the coding of digital images (such as trademark images) or digital audio clips; Meaningless watermarks only correspond to one serial number. The advantage of meaningful watermarking is that if the decoded watermark is destroyed due to attacks or other reasons, people can still confirm whether there is a watermark through visual observation. But for meaningless watermark, if there are some symbol errors in the decoded watermark sequence, we can only determine whether the signal contains watermark through statistical judgment.
-5. According to the purpose
-Different application requirements produce different watermarking technologies. According to the use of watermark, digital watermark can be divided into ticket anti-counterfeiting watermark, copyright protection watermark, tampering prompt watermark and hidden identification watermark.
Ticket anti-counterfeiting watermark is a special kind of watermark, which is mainly used for anti-counterfeiting of printed bills, electronic bills and various certificates. Generally speaking, it is impossible for counterfeit money makers to modify the bill image too much, so signal editing operations such as scale transformation are not considered. However, on the other hand, people must consider the situation that the bill is damaged and the pattern is blurred. Considering the requirement of rapid detection, the digital watermarking algorithm used for bill anti-counterfeiting should not be too complicated.
Copyright mark watermark is a kind of digital watermark that has been studied most at present. Digital works are both goods and knowledge works, and this duality determines that copyright logo watermarking mainly emphasizes concealment and robustness, while the requirements for data are relatively small.
Tampering hint watermark is a fragile watermark, and its purpose is to identify the integrity and authenticity of the original file signal.
-The purpose of covert identification watermark is to hide important marks of confidential data and restrict the use of confidential data by illegal users.
-6. According to the hidden position of the watermark.
According to the hiding position of digital watermark, we can divide it into time (space) domain digital watermark, frequency domain digital watermark, time/frequency domain digital watermark and time/scale domain digital watermark.
Time (space) domain digital watermarking directly superimposes watermark information on the signal space, while frequency domain digital watermarking, time/frequency domain digital watermarking and time/scale domain digital watermarking hide watermarks in DCT transform domain, time/frequency transform domain and wavelet transform domain respectively.
With the development of digital watermarking technology, various watermarking algorithms emerge one after another, and the hiding position of watermark is no longer limited to the above four kinds. It should be said that as long as a signal transformation is formed, it is possible to hide the watermark in its transformation space.
Typical digital watermarking algorithm
In recent years, the research of digital watermarking technology has made great progress. Some typical algorithms are analyzed here. Unless otherwise specified, these algorithms are mainly aimed at image data (some algorithms are also applicable to video and audio data).
-Spatial algorithm The typical watermarking algorithm in this kind of algorithm is to embed information into the least significant bit (LSB) of randomly selected image points, which can ensure that the embedded watermark is invisible. However, due to the use of unimportant pixels in the image, the robustness of the algorithm is poor, and the watermark information is easily destroyed by filtering, image quantization and geometric deformation. Another common method is to use the statistical characteristics of pixels to embed information into the brightness value of pixels.
The-Patchwork algorithm randomly selects n pairs of pixel points (ai, bi), then adds 1 to the brightness value of each ai point and subtracts 1 from the brightness value of each bi point, so that the average brightness of the whole image remains unchanged. Patchwork method can resist JPEG compression, FIR filtering and image cropping by properly adjusting parameters, but the amount of information embedded in this method is limited. In order to embed more watermark information, we can divide the image into blocks and then embed each image block.
-Transform domain algorithm In this kind of algorithm, most watermarking algorithms adopt spread spectrum communication technology. The algorithm calculates the image by discrete cosine transform (DCT), and then superimposes the watermark on the top k coefficients with the largest amplitude in DCT domain (excluding DC component), which is usually the low frequency component of the image. If the first k largest components of DCT coefficients are expressed as D={ di}, i= 1, ..., k, and the watermark is a random real number sequence that obeys Gaussian distribution, W ={ wi}, i= 1, ..., k, then the embedding algorithm of the watermark is di = di (/kloc-0. Then the watermark image I is obtained by inverse transformation with the new coefficients. The decoding function calculates the discrete cosine transform of the original image I and the watermark image I * respectively, extracts the embedded watermark W *, and then carries out correlation test to determine whether the watermark exists. This method can extract a reliable watermark copy even if the watermark image is obviously deformed after some general geometric deformation and signal processing operations. A simple improvement is to embed the watermark into the intermediate frequency component of DCT domain instead of the low frequency component to adjust the contradiction between the robustness and invisibility of the watermark. In addition, the spatial domain data of digital images can be transformed into corresponding frequency domain coefficients by discrete Fourier transform (DFT) or discrete wavelet transform (DWT). Secondly, according to the type of information to be hidden, appropriate coding or deformation is carried out; Thirdly, according to the size of hidden information and its corresponding security target, some types of frequency domain coefficient sequences (such as high frequency, intermediate frequency or low frequency) are selected; Thirdly, certain rules or algorithms are determined, and the previously selected frequency domain coefficient sequence is modified with the corresponding data of the information to be hidden; Finally, the frequency domain coefficients of the digital image are transformed into spatial data by corresponding inverse transformation. The operation of hiding and extracting information in this kind of algorithm is complicated, and the amount of information hidden can not be large, but it has strong anti-attack ability, which is very suitable for digital watermarking technology for copyright protection of digital works.
-compressed domain algorithm The compressed domain digital watermarking system based on JPEG and MPEG standards not only saves a lot of complete decoding and re-encoding, but also has great practical value in digital TV broadcasting and VOD (Video on Demand). Therefore, watermark detection and extraction can also be directly carried out in compressed domain data. This paper introduces a digital watermarking scheme suitable for MPEG-2 compressed video data stream. Although MPEG-2 data stream syntax allows user data to be added to the data stream, this scheme is not suitable for digital watermarking technology, because user data can be simply deleted from the data stream, and at the same time, adding user data to MPEG-2 encoded video data stream will increase the bit rate, making it unsuitable for fixed bandwidth applications, so the key is how to add the watermark signal to the data signal, that is, to the data stream representing video frames. For the input MPEG-2 data stream, it can be divided into three parts: header information, motion vector (for motion compensation) and DCT coded signal block. In this scheme, only the last part of the MPEG-2 data stream is changed. The principle is as follows: firstly, each input Huffman code in a DCT coded data block is decoded and dequantized to obtain the DCT coefficient of the current data block; Secondly, the transform coefficients of the corresponding watermark signal blocks are added to obtain the DCT coefficients of watermark superposition, and then the DCT coefficients are quantized and huffman encoding again. Finally, the bit number n 1 of the new Huffman codeword is compared with the original codeword n0 without watermark coefficient. Only when n 1 is not greater than n0 can the watermark code word be transmitted, otherwise the original code word is transmitted to ensure that the code rate of the video data stream does not increase. This method has a problem worthy of consideration, that is, the introduction of watermark signal is an error signal that causes degradation, and the coding scheme based on motion compensation will spread and accumulate an error. In order to solve this problem, the algorithm adopts a drift compensation scheme to offset the visual distortion caused by the introduction of watermark signal.
-NEC algorithm this algorithm was proposed by Cox and others in NEC laboratory, and it plays an important role in digital watermarking algorithm. The realization method is as follows: firstly, a pseudo-random sequence is generated by taking the key as the seed, which has a Gaussian n (0, 1) distribution, and the key generally consists of the author's identification code and the hash value of the image; Secondly, the image is transformed by DCT, and finally the image is modulated (superimposed) by pseudo-random Gaussian sequence. The algorithm has strong robustness, security and transparency. Because of the special key, it can prevent IBM attacks. The algorithm also puts forward an important principle to enhance the robustness and anti-attack algorithm of watermark, that is, the watermark signal should be embedded in the most important part of the source data and composed of independent and identically distributed random real number sequences, which should have the characteristics of Gaussian distribution n (0, 1).
-Physiological model algorithm Human physiological models include HVS(HumanVisualSystem) and Human auditional system (HAS). This model can be used not only in multimedia data compression system, but also in digital watermarking system. The basic idea of using visual model is to use JND (just perceptible difference) description derived from visual model to determine the maximum intensity of digital watermark signal that can be tolerated by each part of the image, so as to avoid damaging visual quality. That is to say, the visual model is used to determine the modulation mask related to the image, and then it is used to embed the watermark. This method has good transparency and robustness.
Application field of digital watermarking
With the development of digital watermarking technology, the application fields of digital watermarking are also expanding. The basic application fields of digital watermarking are copyright protection, hidden identification, authentication and secure stealth communication.
When digital watermarking is applied to copyright protection, the potential application market lies in e-commerce, online or offline distribution of multimedia content and large-scale broadcasting services. When digital watermarking is used to hide signs, it can be applied to medicine, cartography, digital imaging, digital image monitoring, multimedia indexing and content-based retrieval. In the authentication of digital watermark, the secure invisible communication of digital watermark on ID card, credit card and ATM card will be widely used in national defense and intelligence departments. The rapid development of multimedia technology and the popularity of Internet have brought a series of political, economic, military and cultural problems, and produced many new research hotspots. The following issues that have aroused widespread concern constitute the research background of digital watermarking.
-1. Intellectual property protection of digital works
Copyright protection of digital works (such as computer art, scanned images, digital music, video and 3D animation) is a hot issue at present. Because it is very easy to copy and modify a digital work, and it can be exactly the same as the original, the creator has to use some methods that seriously damage the quality of the work to add copyright marks, and this obviously visible mark is easy to be tampered with.
-"Digital watermarking" uses the principle of data hiding to make the copyright mark invisible or inaudible, which does not damage the original work, but also achieves the purpose of copyright protection. At present, the digital watermarking technology for copyright protection has entered the preliminary practical stage. IBM has provided digital watermarking function in its "Digital Library" software, and Adobe has also integrated Digimarc's digital watermarking plug-in into its famous Photoshop software. But realistically speaking, the technology of digital watermarking products on the market at present is not mature, and it is easy to be destroyed or cracked, and there is still a long way to go before it is really practical.
-2. Anti-counterfeiting of bills in commercial transactions
With the development of high-quality image input/output equipment, especially the appearance of color inkjet, laser printer and high-precision color copier with accuracy exceeding 1200dpi, it becomes easier to forge bills such as money and checks.
On the other hand, in the process of transforming from traditional commerce to e-commerce, there will be a large number of excessive electronic documents, such as scanned images of various paper bills. Even after the network security technology is mature, all kinds of electronic bills still need some non-password authentication methods. Digital watermarking technology can provide invisible authentication marks for various bills, which greatly increases the difficulty of forgery.
-3. Authenticity appraisal of documents
-Information hiding technology can be widely used. As a certificate, everyone needs more than one certificate, which can prove personal identity: ID card, passport, driver's license, pass, etc. Proof of a certain ability includes: various academic certificates, qualification certificates, etc.
At present, China is facing huge business opportunities in the field of document anti-counterfeiting. Due to the lack of effective measures, "counterfeiting", "buying fakes" and "using fakes" have become the norm, which has seriously interfered with the normal economic order and adversely affected the national image. The authenticity of the certificate can be confirmed by watermarking technology, which makes it impossible to copy and copy the certificate.
-4. Hidden identification and tampering tips of audio-visual data
The identification information of data is often more confidential than the data itself, such as the shooting date, latitude and longitude of remote sensing images, etc. Sometimes the data without identification information can't even be used, but it is dangerous to mark these important information directly on the original file. Digital watermarking technology provides a way to hide the logo. Logo information is invisible in the original file and can only be read by a special reader. This method has been adopted by some foreign public remote sensing image databases.
-In addition, data tampering tips are also a very important job. The existing signal splicing and mosaic technology can achieve "grafting" without being known, so how to prevent tampering attacks on image, audio and video data is an important research topic. Tampering hint based on digital watermark is an ideal technical way to solve this problem, and whether audio-visual signals have been tampered with can be judged by hiding the state of watermark.
-5. Covert communication and its countermeasures
The information hiding technology that digital watermarking relies on not only provides a non-password security approach, but also triggers a revolution in information warfare, especially network intelligence warfare, resulting in a series of novel combat methods, which has attracted the attention of many countries.
Network intelligence warfare is an important part of information warfare, and its core content is to use public networks to transmit confidential data. So far, academic research in this field has failed to break through the thinking mode of "file encryption". However, encrypted files are often confusing and easy to attract the attention of attackers. With the wide application of network multimedia technology, using public network for secure communication has become a new idea. By using the redundancy of digital audio-visual signals relative to human vision and hearing, information can be hidden in various time (space) and transform domains, thus realizing covert communication.