Perceptual Feature Guided Color Image Watermarking Research

In recent years,with the popularization and development of network technology and smart devices,the storage,copying and dissemination of digital multimedia content has become easier.The information age not only brings great convenience to people,but also brings a set of multimedia information security issues.As an effective means to protect and authenticate the security of multimedia data,digital watermarking technology plays a greatly important role in security protection problems such as copyright protection and anticounterfeiting authentication.In the era of multimedia information,how to ensure and improve the performance of digital watermarking algorithms has become a key issue for researchers.The contradiction between invisibility and robustness limits the development and application of watermarking algorithms.Therefore,how to coordinate and improve the performance between the two has always been the core problem of the design of a robust watermarking algorithm.In the currently existing solutions,researchers have incorporated visual characteristics into robust algorithms,making watermarked images have achieved good results in terms of fidelity and robustness.With the in-depth study of the human visual system(HVS)and the development of visual characteristic calculation,it is gradually proposed that the characteristics of the HVS can be better represented.Therefore,this article fully considers and uses the perceptual effects of the HVS,and designs visual characteristics that are more in line with the human visual characteristics by perceiving the visual characteristics of the image.To make the pixel changes caused by the watermark embedding more consistent with human visual perception,we design and use the visual feature model to reasonably modulate the modification amount of the pixel,adaptively adjust the modification amount of each pixel.And we realize the update of the pixel value in the spatial domain to achieve the purpose of embedding digital watermark information.The perceptually guided watermark embedding method proposed in this paper realizes the tradeoff between robustness and invisibility,makes the watermarked image more consistent with the perceptual characteristics of the HVS,and effectively improves the visual perception quality of the watermark image.Specifically,the work of this paper is as follows:1.This paper proposes an adaptive quantization step size based on image block classification and a method of using a just noticeable difference(JND)model to guide the amount of pixel modification.Because the human eye perceives the difference in different regions of the image,use the alternating current(AC)in the discrete dosine transform(DCT)block classifies the image block.The adaptive quantization step size is obtained according to the type of the image block.At the same time,this paper uses JND model to sense and guide the modification of pixels,and realizes the embedding of watermark information.2.According to the perceptual difference of HVS to color complexity and visual saliency,this paper proposes a watermarking algorithm based on spatial domain multi feature perceptual guidance.We consider that the human eye has a different sensitivity to the color image content of different mode complexity,also consider the impact of color complexity on the image mode content.It will cause the human eye to produce different visual masks for the complexity of the original image mode.In this paper,we use a color complexity modulation factor and visual salience(VS)model to modulate JND model.In this paper,considering the influence of multiple spatial features on HVS,we propose a spatial domain JND model which combines color complexity and visual saliency.The proposed JND model not only has good visual results,but also has good performance in the proposed perceptual watermark embedding framework.