Application of the hottest digital watermarking sy

The application of digital watermarking system in university digital library the emergence of digital library has greatly alleviated the contradiction between the literature service of University Library and the information needs of readers, greatly facilitated the readers, and made the communication of multimedia information reach the unprecedented depth and breadth. But the following side effects are also very obvious, works infringement is easier, tampering is also more convenient. College digital libraries are in urgent need. A technology that can effectively and safely protect product copyright in order to implement effective copyright protection and information security. With the development of science and technology and the deepening of research, digital watermarking technology has become an effective way to protect the copyright of digital products and has been more and more widely studied and applied. Digital watermarking technology refers to adding specific digital information to digital products (such as graphic, audio-visual and other digital products) according to a certain algorithm, and realizing various functions including copyright protection by detecting and extracting the specific information. Digital watermarking is known as the "last line of defense" against multimedia piracy

1 basic framework of digital watermarking system

digital water as system is generally composed of three parts: water as generation, water as embedding and water as detection and extraction

1) water is generated: in the library digital water is system, the image or icon of the library name is usually used as a digital watermark driven by various resources to mark the ownership of the library collection. In order to enhance security, these information must be processed accordingly (such as encryption, scrambling, etc.) before being used as a digital watermark

2) watermark embedding: the process of embedding the generated watermark information into the original digital product. The watermark embedding should comprehensively consider the transparency and robustness of the digital watermark, and take into account the needs of these two aspects while completing the digital watermark function, which not only ensures the transparency of the digital watermark (the existence of the digital watermark cannot be easily detected), but also has a certain robustness (destroying or deleting the digital watermark will reduce the quality of the watermark carrier or even make it unusable)

3) watermark detection and extraction: the process of detecting the existence of the watermark, recovering the watermark and extracting it. According to the received carrier, learn from the predetermined algorithm and key to detect the digital watermark on the carrier to determine whether there is a digital watermark in the detection carrier. When it is determined that there is a digital watermark, the digital watermark in the carrier is extracted and restored according to the predetermined algorithm and key. The extracted digital watermark can complete the applications of digital library, such as copyright confirmation, product use legitimacy recognition and so on

2 main functions of digital watermarking system in University Library

according to the basic framework of digital watermarking system (as shown in Figure L), the library digital watermarking system should be able to achieve the following local functions:

1) digital watermarking and original carrier preprocessing. The purpose of digital watermarking and the preprocessing of the original carrier is to enhance the security of digital watermarking, which is conducive to the embedding, detection and recovery of digital watermarking. The preprocessing algorithms are usually binary bit serialization, encryption, scrambling and so on

2) embedding, detection and extraction of digital watermarking. Completing the embedding, detection and extraction of digital watermarking is the core of the whole digital watermarking system

3 performance of digital watermarking system in University Library

in order to achieve effective copyright protection, digital products that need to add digital watermarking have the same use function and application value as the original digital products. Digital products will not lose their own function or application value because of the addition of digital watermarking. Digital watermarking has two main characteristics: one is transparency, that is, digital watermarking is imperceptible and does not affect the normal use of digital products. The second is robustness. The digital watermark will not be lost due to signal processing. The digital watermark can still maintain partial integrity and be accurately identified. Transparency and robustness are contradictory. Generally, the more the watermark is embedded, the stronger the strength is, the better the robustness is, but the relative transparency is worse. The performance of digital watermarking system needs to comprehensively consider these two factors and balance the requirements of the two, so as to improve the robustness of digital watermarking as much as possible on the premise of ensuring transparency

3.1 using mature digital watermarking processing technology

in recent years, the research of digital watermarking technology has made great progress, and there have been many mature and effective digital watermarking algorithms. These algorithms have comprehensively considered transparency and robustness in design and practice, and have been constantly improved through a lot of practice. These algorithms can be directly applied to digital watermarking systems, such as spatial algorithm LSB, compressed domain algorithm (based on JPEG and MPEG standards), physiological model algorithm, transform domain algorithm (based on discrete cosine transform algorithm DCT, based on discrete wavelet transform algorithm DWT), etc. Among them, the operation of hiding and extracting information of transform domain algorithm is complex, the amount of hidden information is small, but the anti attack ability is strong, which is very suitable for digital watermarking technology of digital works copyright protection. In the digital watermarking system of digital library, the transform domain algorithm can be used as the main algorithm, which has high robustness while ensuring the transparency and security of digital watermarking

3) anti interference test of digital watermark. The test of digital watermarking is mainly used to measure the ability of digital watermarking to resist attacks (robustness test). The test includes the ability to filter, add noise, JPEG image compression, image rotation, image processing and other attacks

4) calculation and evaluation of relevant indicators of digital watermarking. The related index of digital watermarking is to give quantitative values to the digital watermarking and the embedded carrier to evaluate the performance of digital watermarking. Its evaluation indicators mainly include peak signal-to-noise ratio (PSNR), correlation coefficient (NC), mean square error (MSE), etc

the basic functional structure of the digital watermarking system is shown in Figure 2

the following takes a digital watermarking algorithm in DCT domain as an example to realize the process of embedding, detection and extraction of digital watermarking. Discrete cosine transform (DCT for short) is considered to be the best method to transform speech and image signals. The algorithm of digital watermark embedding using DCT is:

stepl: press the original image to 8 × 8 size, calculate the DCT coefficients of each block respectively, and store the 1-bit information of the digital watermark in each block. Scrambling the watermark image can effectively improve the robustness of the digital watermark

step2: for the DCT coefficients of the i-th block information, where (U1, V1,), and (U2, V2), represent the values of the two coefficients in the block. Read the digital watermark information in turn. When the hidden information is bit L, make (UL, VL) I = (U2, V2) I, otherwise make (UL, VL) I (U2, V2) I

the method of extracting watermark is the reverse operation of embedding. Press 8 for the carrier image with watermark × 8. Divide the blocks and calculate the system of each block. When the coefficient (UL, VL) I = (U2, V2) I, extract bit L, otherwise extract bit 0, access the bit value in turn, and obtain the embedded digital watermark after anti scrambling

3.2 establishment of test system

in digital watermarking system, the performance test of digital watermarking is often consistent with the attack of digital watermarking. Without specific attack methods, the evaluation of digital watermarking system is of no significance. The most common attacks are median filtering, sharpening filtering, mosaic attack, noise adding attack, image rotation, clipping, size change, jppeg compression and blur processing. When testing the digital watermarking system, carry out various attacks on the digital watermarking function and judge the anti attack ability of the digital watermarking. As a reference for the evaluation of the digital watermarking algorithm, the error generated by testing the digital watermarking when it is attacked by the watermark can effectively explain the performance of the digital watermarking system

in the above DCT algorithm, in order to enhance the robustness of the digital watermark, a control quantity alpha is introduced to amplify the difference of the coefficients. Just when writing the digital watermark information, make | (fatigue testing machine advantage analysis UL, VL) I = (U2, V2) I | alpha, so as to ensure that the watermark information can be correctly extracted even if the two coefficients (UL, VL) I and (U2, V2) I are changed (as shown in Figure 3)

Figure 4 shows the carrier image of DNA. Digital water after DCT algorithm when alpha is from 0.1 to o.8. When alpha =0.5, there is little visual difference between the image after adding digital water and the original image. When alpha is 0.5, the traces of DNA and digital water can be clearly seen

Figure 5 shows that when the alpha value is larger, the transparency of the digital watermark is lower, but the ability to resist attacks is stronger, that is, the robustness is higher; On the contrary, the smaller the alpha value, the higher the transparency of the digital watermark and the lower the robustness. In practical applications, we comprehensively consider the robustness of digital watermarking and the requirements of transparent stars, alpha value. Generally, it is appropriate to take 0..5, which not only ensures the transparency of the digital watermark, but also ensures the strong robustness of the digital watermark. At the same time, it can be seen that this algorithm has strong anti-interference for JPEG compression and median filtering, but weak anti-interference for noise and geometric processing. When improving the algorithm, we can strengthen the weak aspects to further improve the facilities in this regard

3.3 establishment of evaluation system

in the performance evaluation of digital watermarking, the transparency of digital watermarking is a very important aspect. The transparency of digital watermarking in digital watermarking system is the evaluation of the original digital products and the digital products embedded with digital watermarking. The evaluation of digital watermarking can be divided into subjective evaluation and objective evaluation

3.3.1 subjective evaluation

subjective evaluation is mainly considered from the perspective of human vision. To make a fair and reasonable evaluation and comparison, we must consider the perceptibility of the watermark in the evaluation process. This evaluation can be based on ITU-R rec.500 quality level. Table L lists the level and corresponding perceptibility and quality. Subjective testing is of practical value to the final quality evaluation and testing, but it is not very useful in the case of research and development

3.3.2 objective evaluation

the biggest defect of subjective evaluation is that the evaluation value varies from person to person. In measuring the transparency of digital watermarking system, it is necessary to provide objective evaluation methods, detect and quantify the difference between the original digital product and the digital product embedded with digital watermarking into a certain value, and use the size of the value as the standard to evaluate the quality of digital watermarking. The main evaluation indexes of digital watermarking system before image are signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR) and correlation coefficient (NC). SNR and PSNR are considered to be the most popular indicators to evaluate signal quality

peak signal-to-noise ratio (PSNR). The peak signal-to-noise ratio measures the ability of embedding watermark. The higher the PSNR value, the better the transparency. If, represents the original image and I'represents the image with watermark, the PSNR calculation formula is:

signal to noise ratio (SNR). If I represents the original image and I 'represents the image with

watermark, the SNR calculation formula is

in specific applications, PSNR is often used to replace SNR. In terms of images. If the PSNR value of the image is more than 20dB, it is considered that the image is subjectively acceptable. The following table lists the values of PSNR, SNR and other indicators of the digital watermark image generated when alpha is different

it can be seen from table 2 that when using the above DCT algorithm to realize the digital watermarking function, when alpha