Compressed Sensing Meets Information Theory

Compressed Sensing Meets Information Theory Google Tech Talk October 7, 2009 ABSTRACT Presented by Dror Baron, Visiting Scientist, Technion - Israel Institute of Technology. Traditional techniques for collecting samples of band-limited analog signals above the Nyquist rate, which is linked to higher frequency analog signal. Compressed Sensing (CS) is the revelation that the optimization routines can be a signal spread by a small number of linear projections of the reconstructed signal based. Techniques as a result, CS-basedCollection and processing of signals scattered at much lower prices. CS offers an enormous potential in applications such as broadband analog-digital conversion, where the Nyquist rate on the technique. Information theory can offer many insights CS, I describe several studies in this direction. First distributed compressed sensing (DCS) offers new distributed algorithms for signal detection using both intra-and inter-signal correlation structures, multi-signalEnsemble. DCS is immediately applicable in sensor networks. So, we use the remarkable success of LDPC codes and graphics algorithms to design low complexity channel CS reconstruction algorithms. Linear measurements play a crucial role not only in compressed sensing, but in disciplines such as finance, where a number of measures needed to estimate the various statistical properties are noisy. In fact, many areas of science and technology to extract information from linear...