Robust Watermark Algorithm For Screen Photograph

Nowadays,information in digital form is easily replicated,transmitted,and stored,which can lead to information leakage and cause security issues.As a widely used information hiding technique,digital watermarking is employed to embed watermark information into the original carrier to prove copyright ownership and embed identity numbers,time,and other information to locate the source of information leakage.However,with the development of digital devices and the significant improvement in their performance,the ways of information leakage have changed.The method of transmitting information by taking a screenshot of the screen makes leakage easier and more concealed.The screen capture process also brings new distortion.Traditional digital watermarking techniques cannot cope with the new challenges brought by these changes,and therefore,it is urgently needed to design robust watermarking algorithms that can resist screen captures to solve the problem of information leakage caused by screen captures.In this paper,we conduct research on the problems of poor robustness,poor invisibility,and less consideration of lighting factors in screen capture watermarking.The main research contents are as follows:(1)To address the problems of inaccurate watermark region positioning,watermark not concentrated in the image’s key areas,poor invisibility,and low accuracy of watermark extraction,this paper proposes a screen capture robust watermarking algorithm based on an Intensity-based Scale-Invariant Feature Transform(I-SIFT)algorithm and Just Noticeable Difference(JND)model.This algorithm first selects feature points using the I-SIFT algorithm and determines the watermark embedding areas based on the density of feature points in the feature region.Secondly,it uses the JND model to determine the intensity of watermark embedding,realizing the association with image texture and brightness.Finally,after embedding the watermark,the coefficients in the watermark embedding neighborhood are optimized and adjusted with constraints.Compared with other algorithms,this algorithm can improve the accuracy of watermark positioning and extraction and has better robustness against screen captures.(2)To address the problems of not considering the influence of lighting regions on watermark embedding,poor invisibility,and poor lighting correction effect in screen capture watermarking,this paper proposes a screen capture robust watermarking algorithm based on two-dimensional gamma function and top-hat transformation.This algorithm first determines the lighting intensity threshold for feature region selection based on the two-dimensional gamma function,then determines the intensity of watermark embedding based on the DCT frequency sensitivity matrix and JPEG compression value,and finally performs lighting correction on the captured image using the two-dimensional gamma function and top-hat transformation.Compared with other algorithms,this algorithm has better invisibility and better robustness against lighting distortion at different screen capture angles.