Robust blind and non-blind detection for digital watermarking

Official URL: http://risc01.sabanciuniv.edu/record=b1169205 (Table of Contents)

Rapid development of Internet has greatly increased the need for creation, storage and distribution of digital multimedia products. This raises, however, security concerns due to digital multimedia products high vulnerability to the illegal copying, distribution, manipulation, and other attacks. To remedy these security issues, in literature, the Digital Watermarking has been developed where the information to be hidden is carried by the watermark signal that is transmitted over the host signal. The capacity of a watermarking system is suffered from much degradation such as channel distortion, filtering, JPEG compression, cropping etc. In addition to those degradations, the host signal interference may even limit the capacity of some systems called blind watermarking systems where host signal is not available to the end-users. To mitigate these sources of errors and increase the capacity of the system, in this thesis, we develop robust detection methods. For this purpose, we devise block normalization based methods for blind watermarking system in Discrete Cosine Transform domain. We also propose the channel reliability estimation based detector for both blind quantization based watermarking system in Discrete Wavelet Transform domain and non-blind watermarking system in Discrete Cosine Transform domain. Simulation results demonstrate that the developed detection methods improve the capacity, bit error rate performance and the robustness of the systems as compared to existing methods against various distortions and attacks.