| With the rapid development of network communication technology, we should attach great importance to the copyright protection, identification and information security of digital products. Digital watermarking technology becomes an important way to solve these problems because of its robustness and invisibility. However, the watermarking technology is not yet widely used in practical application because it is today’s algorithms are not effective against geometric attacks.In this paper a new watermarking algorithm which resists geometric attack is proposed based on digital holography and Fourier-Mellin transform, when QR code is in the process of scanning and decoding. This algorithm has much more balance between robustness and invisibility without affecting the correct identification of QR code. The main research contents are as follows:1. Researching on the ability to resists geometric attacks of the Discrete Fourier Transform and the Log-polar Transformation, respectively. Finally this paper combining both puts forward a new Fourier-Mellin transform to construct a geometric invariant domain, which facilitates the watermark embedded and extraction. Compared with the conventional Fourier-Mellin transform digital watermarking system, the complexity and the calculation of algorithm is greatly reduced.2. Designing a embedding and extraction algorithm framework which can resist geometric attacks. Watermarking is embedded in the middle frequency of the image’s DFT amplitude spectrum rectangular domain, which is converted into a circle using LPT transform. It makes full use of the middle frequency region, greatly improving the embedding quality.3. This section uses conjugate symmetric extension of Fourier CGH methods to generate digital hologram. Because of hologram having strong anti-tear property, it’s embedded in QR code as a watermark, which greatly improves the security of QR codes. Experiments show that this paper’s extracted watermarking has better robustness. |
PDF Full Text Request