Zero-Watermarking in Transform Domain and Quadtree Decomposition for Under Water Images Captured by Robot

Abstract

To protect the copyrights of the digital multimedia in the information security domain, digital watermarking is still a solution. But recently, most of the researchers focused on the zero-watermarking technology where the original image is undisturbed by watermark embedding. In mission-critical surveillance applications (intrusion detection and disaster management scenarios) where the mobile robots are used, it will be a good idea to watermark the images captured by the robots to prevent the copyright infringement. If we watermark robot captured images (RCI) visual quality degradation of the image will be of main concern. So a zero-watermarking scheme using non sub-sampled contourlet transform and quadtree decomposition is proposed for RCI in this article. The approximated sub-band obtained by contourlet transform is undergone quadtree decomposition, to split the sub-band into blocks of different sizes. The sparse features of these blocks and the statistical distribution of the sub-band coefficients are exploited to generate a zero-watermark which is used for ownership authentication. The experimental results show that the proposed method outperforms the existing scheme, and resists most of the geometrical attacks and the signal processing attacks effectively.