| The rapid development of multimedia technology and popularization of Internet has brought a series of problems in the areas of politics, economy, military affairs and culture and yielded many research hotspots, such as safeguarding the intellectual property of numeral works, preventing bond counterfeit in trade, hiding marks and revising suggestion on sound-to-picture data, secret communication, etc. Among these issues, numeral watermark technology has achieved rapid progress and at the same time some typical algorithms have appeared.A complete watermarking system design includes three parts: the generation of the watermark, the embedding of the watermark and the extraction of the watermark. Its important features are transparency, robustness, capacity, security and so on. The initial purpose of designed digital watermarking technology is to protect static images and video. Now it is widely applied to other digital media, such as online music, DVD and digital television. The primary role of digital watermarking is copyright protection, anti-counterfeit, copy protection, distribution monitoring, data authentication, and information hiding. Along with the development of digital watermarking technology, the algorithms of digital watermarking have been continuously improved. It utilized the method of the space domain and transformation domain. Image watermarking technology is the most mature one among the present watermarking technology researches. In space domain algorithm watermarking is added through transforming the image data directly. It superimposed the digital watermarking on the image space domain through certain algorithms. Generally the watermark is embedded into the signal intensity value of the brightness or colorfulness of the watermark image. Space domain watermarking algorithm has the advantage of little impact on the host image, simple algorithm, fast computing, large information embedded, and the effective use of Human Visual System.As the development of technology, color image and video have boomed up in our daily life. The study on digital image should focus on color digital image. Meanwhile, the study on the color digital image watermark also is significant. The algorithm of this thesis is based on the color image as the carrier. RGB(Red, Green, Blue) is the common color space in computers. It generates other colors by additive mix of three primary colors, and it is applied broadly in computer graphics, image system and color television. While CMYK (Cyan, Magenta, Yellow, Black) is generates other colors by subtractive mix of three primary colors, which is applied broadly in print industry. This thesis adopts the RGB format.This thesis presents the design of hierarchical digital watermarking algorithm, and two schemes have been used for the formation and embedding of watermarking information: 1. The scheme based on Torus auto orphism mapping block. 2. The scheme based on bevel isometrics. This algorithm is used for the image content integrality authentication and the image tamper localization and recovery. In the pretreatment part, the original image is disposed by segmentation, and image mapped sequence is formed via the above mentioned two schemes. In the embedding, the original RGB image is decomposed to R,G and B vectors, which then are embedded into watermark separately. The watermark information is generated by the mapped sequence obtained previously, which then is embedded into the 2 bit (LSB) of the original image pixel. The watermark information is composed of authentication watermark bit, parity check code and watermark recovery. Watermark recovery is the chroma average value of subblocks of the corresponding mapping image. In the detection part, the hierarchical idea is used, and the watermark tamper localization and recovery are carried out through three hierarchy. When the tampered image is recovered, pretreatment of tampered image is needed to obtain mapped sequence of image blocks by which the tampered image is recovered. The experimental results show that this algorithm can be used to effective detection, precise locating and content recovery for the local tampering. Between the two schemes the latter generally is more effective for recovery of image tampering.The embedding algorithm adopted in the thesis is the spatial domain LSB algorithm, that is, watermarking information is embedded into the 2 bit (LSB) of the original image pixel, aiming at increase of the effect for the recovery of image tampering. The merits of the algorithm are that the principle is simple, location precision is high, and recovery of image tampering is effective. LSB is the first digital watermarking algorithm proposed by R.G..van Schyndel et al., which is a typical spatial domain information hiddenness algorithm. With this algorithm, information hiddenness is realized through adjusting the lowest digital of the original data, so that detection of information hiddenness is difficult for a common consumer via HVS. The Chinese meaning of LSB is the less important site, that is, watermark image is embedded into the less important site of the carrier image. In an image of 256 intensities, each pixel is composed of eight numbers of binary system, each of which has different action. The highest site represents 10 000 000(128),and the lowest site represents only 1.This gives watermark direct suggestion that data can be embedded through changing the lowest site of an image. The merits of LSB are that the capacity of information embedment is big, the embedment rapidity is rapid, the process of embedment and pick-up is simple. But the disadvantages are that it is fragile, bad robustness, and can not manage the disposal of sieve wave and adding noise. |
PDF Full Text Request