Research On HVS-Based Adaptive Blind Watermarking Algorithm And Halftone Images Watermarking Algorithm

With the rapid development of the computer technology and network technology, the phenomenon of illegal copying and piracy of various multimedia products has become more and more serious, and the raditional encryption technology has become limited to resolve the copyright protection problem, so people have began to seek the rather method to resolve the problem of copyright protection. Digital watermarking technology has become a new network safety technology in recent years, some digital information, such as user ID, product symbol and significative letter and so on, can be embedded into the multimedia data in order to protect data, which don't affect the value of original content and can't be seen on the image. The watermark can be extracted while needing in order to achieve copyright protection. The applications of digital watermarking are very broad, which can achieve the functions of broadcast monitoring, distinguish of owner, proof of ownership, operation tracking, content verification, copy control and equipment control, etc.Currently, many traditional media contnet are changed toward digital, the carrier and the manner of civilization transmitting are also changed greatly, but the demand of paper has taken on the increasing trend along with the increase of electronic products. That is to say, during the broad application of digital multimedia(audio and video frequency), most image information will exist and transmit in the form of printing in a long time, such as book, magazine, newspapers, printer output and fax, and so on, which all belong to halftone images. Therefore, we not only need to research that some digital information are embedded into multimedia products, but also need to research how to embed the invisible mark(such as corporation symbol, ownership information, production date, copyright protection and authentication information, and so on)to the halftone images to achieve copyright protection. Although many algorithms have been proposed for images to embed watermark, most of them are aiming at gray image or color image, which aren't applied to halftong image directly because of characteristics of halftone image itself. so we need to seek suited watermark algorithms for halftone image. Anyway, the research of digital watermarking technology have very important significance and practicality value for thecopyright protections of multimedia information and halftone products, the illegal copying and modifying of products.The paper is divided into six chapters, the detailed arrangement as following: In Chapter One, introduce the study background of the subject, and introduce the origin and the development status of the digital watermarking, the development status of halftone image watermarking, the main application of the digital watermarking and the existed problem of current research detailedly.In Chapter Two, the basic knowledge of digital watermarking and image halftoning disposal are introduced, including the digital watermarking theory framework, definition, characteristics, classification and evaluating standard. In this chapter, we also introduce some methods of halftoning disposal in commom use, Fourier Transform Algorithm, Fast Fourier Transform Algorithm, the definition and characteristics of Discrete Cosine Transform, and so on.In Chapter Three, the classical digital image watermarking algorithms are introduced, including gray image or color image and halftone image. For gray image and color image, the watermarking algorithms are sorted two classes: spatial domain and transformation domain watermarking algorithm, and for halftone image, the watermarking algorithms are also sorted two classes: the watermark is embedded during halftoning and the watermark is embedded after halftoning. This chapter introduces the two kind algorithms detailly, tests part algorithms and analyzes the virtues and defects of the two kind algorithms.In Chapter Four, an adaptive blind watermarking algorithm based on human visual system is proposed, by quantifying DC components adaptively and adjusting DCT mid-frequency coefficients position to embed the watermark, which is a blind watermarking algorithm. Experimental results show that with its imperceptibility, the algorithm is effective and robust to common image processing and some geometric attacks, such as JPEG lossy compression, scaling, adding noise, rotating, filtering, and so on.In Chapter Five, a kernel switch digital watermarking algorithm for halftone image, based on pels, is proposed in this chapter. By judgeing the compoundings of corresponding pels of halftone image without watermark and the watermarked halftone image, we use different error diffusion kernel to dispose the original image to embed the watermark, and introduce two simple methods to detect the watermark. Simulation results show that the detection algorithm can detect the watermark efficiently, when the halftone image with the watermark is severelystained or scratched. Aiming at low robustness of the existed algorithm, Soo-Chang Pei and Jing-Ming Guo proposed the kernel switch halftone image watermarking algorithm based on block, which we simulate, and though selecting coefficients renewedly, the embedding capacity is increased to four times of the original, experiment results show that without great visual effect, the clear watermark can be extracted. But two algorithms have a long distance to application.In Chapter Six, we summarize the whole paper, and show out some questions to be resolved in the digital watermarking in the future.The main innovative work of this paper focus on the following two aspects.(1) Aiming at the defects of the existing watermarking algorithms in the discrete cosine transform domain, from the selection of digital watermark information, the characteristic of coefficients after transforming, the embedding and extracting algorithm of the watermark to consider, an adaptive blind watermarking algorithm based on human visual system is proposed. We choose the binary image as the watermark, by quantifying DC components adaptively and adjusting DCT mid-frequency coefficients position to embed the watermark, which is a blind watermarking algorithm. Experimental results show that the algorithm is effective and robust to common image processing and some geometric attacks, such as JPEG lossy compression, scaling, adding noise, rotating, filtering, and so on, and the algorithm realizes the better balance between imperceptibility, robustness and watermark capacity.(2) Aiming at halftone image, a kernel switch digital watermarking algorithm for halftone image, based on pels, is proposed. By judgeing the compoundings of corresponding pels of halftone image without watermark and the watermarked halftone image, we use different error diffusion kernel to dispose the original image to embed the watermark, and introduce two simple methods to detect the watermark. Simulation results show that the detection algorithm can detect the watermark efficiently, when the halftone image with the watermark is severely stained or scratched. But the algorithm is not enough robust to the attack of printing and scanning, so it has a long distance to application.