Multimedia distribution

The rapid improvement of the Internet technology and the invention of new media distribution channels have created a new area of research. The need to manipulate, enhance and protect the distributed digital information has become a necessity. In our thesis we propose possible solutions to practical problems that arise in the new digital information distribution channels. In particular, we concentrate on personalized distribution services and copyright protection.; In the first part of the thesis we concentrate on the personalized music delivery issue, as an example for the personalized service. We propose two solutions for constructing a sequence of tracks that satisfies user requirements and at the same time optimally exploits music catalogs. The proposed systems allow online users to construct a sequence of songs that satisfy their listening taste and at the same time achieve producers goals. We use both vector space concepts and the network flow model to formulate the constrained sequence retrieval problem as a binary integer linear program (BILP). Our experiments demonstrate the validity of the proposed model to the original problem in reducing the search space and producing their exact solutions.; Protecting media information, from being copied or tampered with, is of vital significance to producers. This is due to the fact that perfect copying and reproduction of digital information is absolutely possible and therefore a huge revenue loss is also likely. In our thesis we propose algorithms for protecting the digital media via the data hiding and watermarking technology. Due to contract constraints, we focus on images as an example for the digital information. The main focus in our research is to combat geometrical distortions. Namely, scaling, rotation and flipping attacks. These could be fatal to the watermarking techniques. We propose three possible solutions for such a problem.