Computational Methods for Bayesian Object Recognition: Theory and Application

Computational Methods for Bayesian Object Recognition: Theory and Application

von Russell J. Bowater
Taschenbuch - 9783838351995
79,00 €
  • Versandkostenfrei
  • Hinweis: Print on Demand. Lieferbar innerhalb von 7 bis 10 Tagen
  • inkl. MwSt. & Versandkosten (innerhalb Deutschlands)
  • Hinweis: Print on Demand. Lieferbar innerhalb von 7 bis 10 Tagen
  • inkl. MwSt. & Versandkosten (innerhalb Deutschlands)

Autorenfreundlich Bücher kaufen?!

Beschreibung

During the last few years, statistical methods for the automated recognition of objects within images have become increasingly popular. This book is aimed at providing the reader with a comprehensive understanding of how computational methods, and in particular sampling methods, can be used to facilitate the Bayesian approach to object recognition. A detailed overview is presented of both the general use of Markov chain Monte Carlo (MCMC) methods in sampling and the use of template-based models in statistical image analysis. Special attention is given to reversible jump MCMC methodology which in terms of object recognition allows the introduction and removal of parameters, e.g. whole objects, from the global image model. Clear mathematical definitions of methods discussed and their theoretical properties are provided throughout the book. In addition, a substantial part of the book is devoted to practical applications of these methods. The two core examples that are examined in detail are a toy problem that concerns the imaging of an unknown number of randomly arranged discs and a real world problem that concerns the imaging of microscopic algae filaments.

Details

Verlag LAP Lambert Academic Publishing
Ersterscheinung November 2012
Maße 220 mm x 150 mm x 15 mm
Gewicht 386 Gramm
Format Taschenbuch
ISBN-13 9783838351995
Auflage Nicht bekannt
Seiten 248

Schlagwörter