Research On Non-Digital Domain Image Watermarking Algorithm

Research on print and scan,print and photograph is critical to the future of digital watermarking.The need of connectivity between paper and the digital world is increasing.If the solution to this problem is found,paper,including books,newspapers,magazines,can be protected by digital watermarking,which is undoubtedly a revolutionary technology for copyright protection.With the development and popularity of printers and scanners,it has become easier to convert electronic documents or images from electronic to printed versions.The popularity of shooting equipment has made it easier and uncontrolled to take photos of images or documents.Therefore,it becomes necessary to design a robust watermarking scheme robust to print and scan or print and photograph,display and shooting.This paper proposes three watermarking algorithms in non-digital domain.The non-digital domain watermarking algorithm is an image watermarking algorithm applied to print scanning or shooting,in which the image is converted from digital signal into analog signal and then converted into digital signal again.A print-cam robust watermarking algorithm based on self-convolution is proposed.In the watermark embedding process,first convert RGB color image into CIE Lab color space,and embed the periodic watermark in two color channels of Lab color space.The embedding strength of the watermark is adjusted by noise visible function and the intensity coefficient.In the watermark extraction process,compute self-convolution in order to achieve watermark synchronization and to resist the effects of geometric transformations such as rotation,translation and scaling.Finally,the normalized correlation coefficient between the extracted watermark and the original watermark is calculated to determine whether the watermark exists.Experimental results show that the algorithm realize real-time detection of watermarks without the host image.A traceable watermarking algorithm is proposed for binary document images.Different document versions are formed by embedding different watermark patterns into the same document images to be distributed.Display the document on the screen,take a picture with the camera to get a document image,and extract the watermark from the captured photo.According to the different watermark content extracted,different versions of the document are distinguished,thereby document traceability are achieved.In order to improve watermark detection rate,the watermark detection operation is performed in different areas of the same image.The experimental results show that the success rate of watermark detection is high when the embedding strength is greater than a certain threshold.A multi-bit watermarking algorithm for screen shot images is proposed.The 18-bit binary sequence after error encoding is corresponding to an angle of 0-180 degrees.The pseudo-random image is rotated by the corresponding angle,and then a centrally symmetric image is constructed,and the image is periodically tiled and embedded as a watermark.Since the image may be cropped at the time of shooting,four image blocks with the same watermark information are located when the watermark is extracted,and each block is further divided into blocks to solve the autocorrelation.According to the distribution angle of local peaks of autocorrelation function,three corresponding binary bits are known.Then integrate the four sets of watermark results to get the final watermark information.