CMStatistics 2017: Start Registration
View Submission - CMStatistics
B1262
Title: Maximum likelihood estimation for general models in size and shape space Authors:  Ian Dryden - University of Nottingham (United Kingdom)
Alfred Kume - University of Kent (United Kingdom) [presenting]
Phillip Paine - University of Nottingham (United Kingdom)
Andrew Wood - The University of Nottingham (United Kingdom)
Abstract: Inference in shape analysis is related to problems where the invariance of rotations and translations (and possibly scaling) of data objects is required. If some multivariate distribution is assumed in the landmarks space the estimation is not straightforward. We will represent a general approach for maximum likelihood estimation for such models in size-and-shape spaces, where a general time dependence and covariance among landmarks is allowed. In particular, while for the 2-d case we can use the Bessel functions to carry out the MLE procedure, in the 3-d case we make use of the recent numerical methods for calculating the quantities of interest. Real and simulated data are used to illustrate the proposed method.