Research Of Digital Image Watermarking Techniques Based On AQIM

The rapid development of computer technology makes the digital multimedia technology step onto the stage of a full expansion,and makes the dissemination, replication and storage of the Multimedia information simpler,easier and popularer.Because of the internet, both the material life and the spiritual cultural life of mankind have been greatly improved. One can easily find documentation they need to download, alter, store and disseminate from the online, which appeared in the copyright infringement problems, leading to the copyright protection of digital products and digital transmission of information security issues. How to effectively ensure information security and digital multimedia legal rights of copyright holders may be concerned about the industry and academia and the major problems solved. Then digital watermarking technology in this context is created. Digital watermarking technology is copyrighted information (such as text, images, video), by a certain algorithm, embedded into the original carrier of information, if there is copyright dispute, the copyright information by extracting them, that is, we talk about Watermark information, to safeguard the rights of copyright holders. Through the efforts of several generations of researchers,the watermaking has become the main form of the copyright protection of digital products, based on previous studies, this paper poses an algorithm of image watermaking based on AQIM,which is simulated by matlab . The main contents include the following:First, the basic principles of digital watermarking technology, characteristics, classification, and common attacks were simply described and summarized, often on the current simple digital watermarking system used in the performance evaluation criteria are introduced and analyzed, and simple lists The current range of digital watermarking technology for the future of the watermarking technology provides a theoretical basis for further research.On the basis of the digital watermarking, we introduced two digital watermarking algorithm ,those are the traditional spread spectrum (SS) which is a commonly used algorithm,and Quantization Index Modulation (QIM) which disappeared in recent years,then the quantitative index modulation algorithm was Highlightly introduced.on this basis ,a new watermaking algorithm which is against copy attacks was proposed .In this algorithm,we first extracted the image characteristics as the digital signature from the original carrier image ,then in the DCT domain ,the digital signature and the watermaking information were embedded into the mid-low–ferquency coefficients of the carrier image at the same time.At last we carried the simulation experiment using malab.The experiment rusults shew that, the algorithm maintain the original image in good image quality ,and the same time it was robust to the common attacks,such as the median filtering, cropping, brightness adjustment, salt and pepper noise attacks,however, it was poor performance for the for Gaussian noise and JPEG compression attacks.This scheme is hoped to be more improved.Because of the embedding of the digital signature as well as the watermake ,it has a good effect of anti-copy attack. This algorithm is quantified using a fixed quantization step, so there are some attacks on the robustness of the results is not very good,and the extraction and detection was explicited which made this algorithm limitated.Based on Quantization index modulation (QIM),we proposed the concept of adaptive quantization index modulation (AQIM), and further proposed an adaptive quantization index modulation algorithm based on the HVS model.After the the introduction of nature of the HVS model,refering the brightness sensitivity and frequency sensitivity,we devided the original image by 8*8 in the DCT domain,and made the mean of the DC coefficient as a quantization step for each, then made the watermark embedded into the mid-low frequency coefficients. DCT DC coefficient of the image characterized most of the information, when the embedding position changes,the step can be adaptively modulated according to the regulation linear changes of the brightness of earch block. The choice of low-frequency frequency of each block were where the value of the frequency matrix was between 0 and 0.25 . coincidentally they were the first 2-15 datas.in the DCT domain by zig-zag arrangement .We extracted and detect the wartermake from the wartermarking image by the minimum distance detection and the joint decision method to achieve blind watermark extraction..At last we simulated the algorithm using matlab, the experimental results show that the algorithm can maintain the original image good visual effect and good robustness on the common attacks. By comparison with other literature, the effects of the algorithm against JPEG compression and multiplicative brightness adjustments are particularly good , for the other attacks, such as Gaussian noise, salt and pepper noise, cropping, brightness adjustment additive, scale ,the effects are better than other comparative literatures. However,the effcte of the resistance on the median filtering window is common., in this respect, the algorithm has to be improved and further researched.The speed of the development of the digital media nowadays is beyond people's imagination,which brings the material and spiritual superiority impossible to measure. but it also brings a huge threat. After so many years of development, digital watermarking technology has been recognized by the academic community and the great concern and has made a lot of results.However,with the development of the watermarking technology, the attacks on watermark are also constantly upgrading and new attacks are generating.It has so many problems to be solved and there are new research topics not to be touched. Now, the evaluation of digital watermarking technology is still with the measure Perfect,for example, the watermark embedding, detection / extraction, watermark, etc. are the lack of standardized testing requirements,and visual evaluation of digital watermarking is just a subjective concept, there is no specific quantitative analysis. In the human visual system and the evaluation of digital watermarking technology, a clear objective criteria compare match should be established.About quantization index modulation(QIM), the adaptive quantization step is a good way to improve the watermark embedding,but the method to modulate the adaptive quantization step is monotonous. How to choose the adaptive variable and how to achieve the adaptive modulate in the AQIM require deep discussion.Finally,the paper sumerized the digital watermarking technology and the development of this field were prospected. Hope the time will soon be able to see a great leap watermarking technology and development.