A Robust Digital Watermarking Algorithm Based On Image Content Sensibility Analysis

Recently, the explosive growth of the Internet has increased multimedia services, such as electronic commerce, pay-per-view, video-on-demand, electronic newspapers, and peer-to-peer media sharing. As a result, multimedia data can be obtained quickly over high speed network connections. However, authors, publishers, owners and providers of multimedia data are reluctant to grant the distribution of their documents in a networked environment because the ease of intercepting, copying and redistributing electrical data in their exact original form encourages copyright violation. Therefore, it is crucial for the future development of networked multimedia systems that robust methods are developed to protect the intellectual property right of data owners against unauthorized copying and redistribution of the material made available on the network.A possible solution consists of marking multimedia work to allow their spreading to be tracked: digital watermarking presents a viable solution to the above problem, since it is possible to identify the source, author, creator, owner, distributor or authorized consumer of a document, by means of the identification code. The watermark is permanently embedded into digital data for copyright protection and for checking if the data has been corrupted. One of the important issues of watermarking is how to compromise between the invisibility and the robustness. First, the watermark should be embedded in an invisible way to avoid degrading of perceptual quality. Second, the watermark should be robust against watermark attacks which are applied to image content. These attacks include but are not limited to lossy image compression, filtering, noise-adding, and geometrical modification.In this paper, we present a novel digital image watermarking algorithm based on content sensitivity analysis. A new method which successfully combines the characteristics of singular values and variances is developed to measure the activity of sub-blocks of the host image. The application of the proposed embedding scheme makes the watermark capable of blind-extraction. Two new means named ICSP-PSNR is presented to evaluate host image's quality. Another significant innovation of our algorithm is that we propose a different means of watermark recovery based on neural network, which contributes to accurate data tracing. Simulation results show that our method performs better than traditional methods and fulfills a good compromise between the robustness and transparency.