Adaptive Quantization Watermarking Based On Visual Cognition Mechanism

In the information era,information communication and sharing have reached unprecedented width and depth with the universal application of Internet.The forms of information also become more abundant,from the traditional paper to digital multimedia(digital image,audio,video etc).Work and life tend to be digitized and networked.The appearance of digital multimedia has brought a lot of convenience.Information storage,copying.modification and dissemination become simple and quick.However,the numbers of subsequent piracy,infringement and malicious tampering are increasing.Rights and interests of the legitimate users have been seriously damaged.The appearance of digital watermarking technology provides an effective solution to this problem.Watermarking is an important category in the field of data hiding.We can realize the protection of multimedia content and the authentication of copyright by embedding information in multimedia data.It has broad application prospects.So it has received wide attention from the society.As a classical algorithm,Quantization Index Modulation(QIM)algorithm shows good performance in terms of efficiency and computational complexity.However,the limitation of this method is that the uniform quantizer is used for all carrier coefficients,and the masking property of human vision is neglected.In order to improve the watermarked image quality and robustness of the watermarking,many researchers have introduced the human visual characteristics to the watermarking algorithm.In this paper,we proposed STDM watermarking algorithm using a robust luminance-based JND model.whose formula of mean brightness is modified.We take the visual attention into account further.Make the visual attention factor modify the JND model,which improves the image quality and the robustness of watermarking.The main innovation and research achievements include:1.Propose STDM watermarking algorithm using luminance-based JND model(STDM-LmJND algorithm).Use the luminance-based JND model to calculate the visual redundant information for the procedure of watermarking embedding and detection.By modifying the formula of mean brightness,the robustness of the JND model is guaranteed,which makes the watermarking detection accurate.Experimental results show that the proposed algorithm in this paper improves the visual quality of the watermarked image.and has a high robustness against Gauss noise,salt and pepper noise,JPEG compression.2.Propose STDM-LmJND Algorithm based on Visual Attention.Based on the STDM-LmJND algorithm,taking the characteristics of human visual attention into account,and using the Saliency Detection by combining Simple Priors(SDSP)method to get the "attention" and "non-attention" region of the image.According to the result.we use different visual attention factors to modify JND coefficients in different regions,in order to get the visual redundancy information closer to the human visual characteristics.the quantization step size will be modified according to the new JND coefficients.Experimental results show that the proposed algorithm improved the robustness and image quality futher.3.Design and implement an Internet information security service platform based on Android system.We have designed and implemented an application for the protection of the image content in the Android platform.At the same time call WeChat,QQ,micro-blog and other social applications interface to facilitate the sharing and spreading of the protected image.The application use the classic STDM algorithm,because many excellent algorithms are based on STDM.So the implementation of the STDM algorithm is helpful to the extension and update of the later application.Mobile terminal is widely used as a platform for information dissemination.It has a direct effect on the security protection of Internet resources.And it has great significance on the popularization and application of digital watermarking technology.