Watermark: from paper texture to digital media

Paper watermarks were invented in the 13th century in Italy for the identification of the workers, the concept continued to evolve in the digital era, to provide authentication and protection for the digital media. In this project, we are interested in digitising and extracting the graphical design of the watermark in order to create a digital version of it using various image processing operations. The digital representation can be embedded into the digital image of the content (the document) to continue its original purpose to protect the content. This project differs from other methods of watermark extraction because it uses back-lighting techniques rather than typical radioactive materials. Hence, it is safer and easier to use. Also, the developed system deals with papers with foreground interference and paper defects, while other methods deal with plain papers without any foreground interference. The software prototype consists of a set of image processing operations which enhance, filter, and extract the watermark shape, and to automatically convert it into a graphical object that can be represented in a vector format. The main purpose of this system is to enable a further level of document imaging for preservation, so that it will be accessible in a digital version. It will also allow the collection of a graphical watermark database which will provide the necessary data for authentication and copyright protection of digital media, and other domain specific studies of the watermark.