摘要:Fitting geometric shapes to observed data points (images) is a popular task in computer vision and modern statistics (errors-in-variables regression). We investigate the problem of existence of the best fit using geometric and probabilistic approaches.