Research On Semi-fragile Watermarking Technology Based On JPEG

With the quick development of the Internet and multimedia techniques, the unauthorized use and vicious tamper of multimedia caused by the openness of the network which make the interest of digital product owner and digital product users damaged. So it is urgent need of a effective instrumentlity to protect the copyright and intellectual property rights of digital product. The traditional encypytion techniques could not provide protect for the data which make it unable to fundamentally solve the problem of unauthorized use and tamper. This is because the data after the crack or decrypt will be very easy to be modified, copied and transmitted.Watermarking technology is developed in this situation. The watermarking technology is a kind of technology that embeds some information into digital carrier sightlessly, the embeded information is called watermark. Only by some detect arithmetic the existing of watermark could be detected. The carrier of watermark includes image, audio frequency, video and software etc. Now, the research on watermarking mainly takes image as carrier. The embed area of digital image watermarking has been developed from airspace to frequency domain. In the manner of watermarking embed, the embedding technology about watermarking has been transformed from fragile watermarking to semi-fragile watermarking combined with robust watermarking.Authentication watermarking is an important area among current watermarking fields, the image authentication technology based on watermark includes fragile watermarking and semi-fragile watermarking. This two kinds of technology play an important role in the field of authenticity and integrality protecting about digital product. During authentication, fragile watermarking will could not pass the detection no matter the image had been experienced normal compression or vicious tamper, which means any change of the image is not allowed, during the progress of authentication, even one byte has altered also could cause the progress result in failure. This is not practical with the increase of internet transfers. Further, with the combination of robust watermark make the semi-fragile watermark has the ability to resistance some normal treatment and also could pass the integrity authentication. These features make the semi-fragile watermark meets the current needs of apply, so this technology was concerned by more and more researchers naturally. Because image has some redundancies, the image usually needs compression before transmission or use. So, make the semi-fragile watermark have the ability to resistance compression must be resolved. Now, many scholars at home and abroad has researched in the field of semi-fragile and has obtained many research results, but most of the existing algorithms only focus on the ability of resistance to compression,noise resistance, tampering and recovery which make these algorithms could not be widely used in practice. JPEG lossy compression is widely used in compression of image, especially in the area of website and digital cameras. The JPEG has become the actual standard for image compression. So it is a great need for the research of semi-fragile watermarking combination with JPEG lossy compression.JPEG is based on the DCT which removal high-frequency information by quantitative to the DCT coefficients. After the progress of quantitative, the low-frequency information which important for the image was preserved. During encoding,because the high-frequency information has been removed, only the low-frequency information need to be encoded. Compared with the unquantitatived coefficients, quantitative significantly reduced the amount of data used to coding an image which can achieve the purpose of the compression. This article proposed a kind of semi-fragile watermark algorithm which could used to content certification after acombined those features of JPEG..This semi-fragile watermarking algorithm is mainly based on the DCT and DWT, and fully use the idea of airspace. Using the feature of the highest four bits of a gray-scale image is important to visual impact. Only the highest four bits of watermarking image are selected as the watermarking information. In the field of watermarking security,use Arnold algorithm to scrambling and encruption the image. The select of original image use the LL3 which generated by DWT. The image of LL3 is the approximate image of original image and could resistance JPEG compression.During the watermark generation, we should pick up the LL3 of original image by DWT in the first setp. The lowest four bit has a little impact on visual quality, so we should set the lowest four bit of the LL3 set to zero if we need to reduce the amount of watermark. Then we will take the image of LL3 as watermark after Arnold scrambling. At the time of embed watermark, we need use the Arnold scrambling to handle the original image,then form a matrix with size of 8×8 by the DCT conversion. After these steps, we need to embed the watermark on the DCT coefficient. The watermark image will be formed after IDCT and anti Arnold to the DCT matrix.The extract of the new watermark needn't to use the original image or the old watermatermark, the alorithm is a kind of completely blind digital alorithm. The process of extracting watermark is the inverse process of the embedding watermark, with that we can get a new watermark image through transfroming the old watermark image with DWT. One can get the orginal watermark image through prikey k with Arnold disturbing and transform towards the watermark image. By comparing the extract of the original watermark with current watermark image, the judge of modification and the positioning of can be done. At the time of the recovery of modification, the watermark image can be zoomed in to recovery the part that has been modified. Because the watermark extracted from the original map is a low resolution approximation image of the original image.In order to verify the reasonability of the algorithm, we take Matlab as the platform for experiments. It can be verified by the experiment that the algorithm advanced by this thesis can resist the operation of JEPG compression and modification relatively. The algorim has the perferable ability of inspecting and locating the modification. In the process of inspection, the algorithm needn't to use the original image and the original watermark to inspecting and locating, which is completely blind digital algorithm. So the algorithm is a useful algorithm in the field of the digital watermarking. At the same time, the experimental results compared with the referenced literature are given. By comparing the results, the performance of the algorithm are better or similar to the algorithm referenced in the literature. By comparing the experimental data we can know that the algorithm proposed in this article is feasible and effective.