Optimal Consensus Maximization for Robust Estimation of Multi-view Geometry in Computer Vision

old_uid7365
titleOptimal Consensus Maximization for Robust Estimation of Multi-view Geometry in Computer Vision
start_date2009/09/24
schedule11h-13h
onlineno
summaryMany Multi-View-Geometry problems in Computer Vision eventually boil down to a robust estimation problem where the task is to estimate a parametric model form a set of noisy measurements often containing outliers (gross error) as well. To solve such a robust estimation problem, vision researchers have adopted various techniques including the well-known RANSAC.
responsibles<not specified>