Research On Adaptive 3D Color Image Watermarking Algorithm Based On Improved Visual Model

Nowadays, the digital information society has arrived in our life.With the matures of information and technology science,the network interconnect technology grows rapidly and changes our living environment all the times. Communication or business between people is not only by phone or face to face. In other words,it is not constrained by time and geography.The development of network interconnection has broke this limitation.Real-time communication and cross-regional transactions can be realized anytime.The Internet also provides download service of multimedia data such as video, audio and text.All these facilitate our lives.However, due to the open nature of the Internet,there are many Internet crimes like a large number of pirated books, videos and pictures.This not only causes a large number of Internet data redundancy, but also against the legitimate rights.More seriously it also disrupts the order of economic.In order to solve the problems, the law has introduced a lot of rules and regulations.But we also need a method to prove its ownership---digital watermarking.Digital watermarking means hiding information in a digital information, playing the roles of marking personal identification information and symbolizing status.This digital information is difficult to be founded but it can be detected by some software.It can also be extracted by the key.There are many types of watermark information, such as text, picture, the DNA sequence information of the owner and fingerprint information.These related digital watermark information can identify the uniqueness of the protected object.Embedding algorithm can ensure that the visual perception is not affected by mass objects to protect copyright.From the point of the copyright protection,this paper we put forward a robust watermarking scheme based on visual model which in the light of the human visual system.The paper introduces the background,the history and the ideas of further research at the beginning.The paper also introduces the classification and it's characteristic,and has carreied on the detailed elaboration to the embedding principle, the quality of assessment methods and the attack method after simulation in the view of watermarking,then.And the general watering method is summarized,which provides the theoretical support for the following research.The introduction of Watson visual model foreshadows the discussion combined with the human visual model.The improved Watson visual model is combined with the modulation algorithm of the quantization index modulation,the purpose of which is to solve the problem that the contrast masking(GAP) does not change with the increase of amplitude and the linear change in the visual model. Algorithm proposed in this paper is to calculate the carrier image luminance masking,contrast masking and perceived quality based on the improved Watson model,and adaptivelly setting the quantization step,eventually adjust the intensity of watermark needs to be embedded automatically according to previously proposed conversion formula,in order to satisfy invisbility characteristics,maximum improve the carrier capacity and the robustness of watermark information.A wide variety of attacks have increased due to the large number of watermarking algorithm currently.In order to improve the robustness of the watermark information against the collusion attack,the algorithm is proved.Aiming at the characteristic that the collusion attack has great influence on the quantitative step growth, we add a module which can determine the quantization step according to the set of predefined perceptual quality range. Combined with the conversion formula to get the quantization step, and discard the minimum value.The simulation data show that the algorithm proposed in this paper has a good robustness for the Gaussian noise attack, salt and pepper noise attack, brightness change, scaling,low pass filtering, and can balance the invisibility of the watermark information. The contrast experiment shows that the improved algorithm is superior to the improved algorithm. Overall, the algorithm proposed in this paper has made great progress in improving the robustness of watermarking.