Advanced Linear Identification Techniques for Signal Processing and Digital Video Broadcasting

Abstract

Linear identification technique is to linearly embed a piece of unique information into digital media data for the purpose of satisfying specific demands such as identification, annotation, and copyright, etc. We need to consider the quantity and the quality of identification data to be embedded as well as the corresponding interference to the original subject signal. However, there exist no generalized computationally-efficient optimization techniques for linear identification up to now. Therefore, in this dissertation work, we try to theoretically investigate the advanced linear identification techniques and combat the tradeoff problems between the quality of the embedded identification data and the quality of the subject signal. Two particular signal processing and telecommunication applications, namely transmitter identification and digital watermarking, will be exploited in this work. We propose a novel optimization paradigm for both digital terrestrial television (DTV) systems and multiple digital watermarking systems to maximize the overall signal-to-interference-plus-noise ratio (SINR) over both identification and subject signals. The new theories and practice related to pseudo random sequences, extended arithmetic-geometric mean inequality, and constrained overall system performance are also presented in this dissertation.