Kuruoglu E. E., Knuth K. Special Issue on Bayesian Source Separation. In: Digital Signal Processing. Editorial, vol. 17 (5) pp. 855 - 857. Ercan Engin Kuruoglu, Kevin Knuth (eds.). Elsevier, 2007. |

Abstract (English) |
The signal processing problem known as source separation has rapidly grown in the last decade from being viewed as something of a curiosity to becoming a fundamental signal processing problem that is ubiquitous across scientific disciplines. Source separation problems are characterized by a set of sources that either emit or modulate signals that propagate to one or several detectors. The signal processing goal is to "separate" the recorded signals into the set of "source" signals. This basic situation appears in a wide variety of contexts ranging from the more traditional mixing of sound signals as in the Cocktail Party Problem to mixtures of source signals of spatial extent in images, such as in astrophysical applications where the objects of study are optically thin (transparent) or magnetic resonance imaging where the effects of several distinct processes are superimposed. At this point in time, source separation has been widely studied, resulting in an extensive array of source separation algorithms. Much of the effort has gone into developing what are known as blind source separation algorithms, referring to the fact that these algorithms are provided with a minimum amount of information about the nature of the recorded signals. These blind algorithms are extremely useful, as they are specifically designed to be generally applicable to a wide array of problems. Many of these techniques work well in a wide variety of situations, including those where noise is an issue. | |

URL: | http://www.sciencedirect.com/science?_ob=PublicationURL&_tockey=%23TOC%236768%232007%23999829994%23664896%23FLP%23&_cdi=6768&_pubType=J&_auth=y&_acct=C000061181&_version=1&_urlVersion=0&_userid=5511047&md5=4bf3d74371b8d9d20821203c1f8e6d5a | |

Subject | Bayesian source separation Independent component analysis G.3 Probability and Statistics. Probabilistic algorithms (including Monte Carlo) 62C10 Bayesian problems; characterization of Bayes procedures 62F15 Bayesian inference |

1) Download Document PDF |

Open access Restricted Private