Statistical 2D and 3D shape analysis using Non-Euclidean Metrics

Rasmus Larsen, Klaus Baggesen Hilger, Mark Christoph Wrobel

AbstractWe address the problem of extracting meaningful, uncorrelated biological
modes of variation from
tangent space shape coordinates in 2D and 3D using non-Euclidean metrics. We
adapt the maximum autocorrelation factor analysis and the minimum noise
fraction transform to shape decomposition. Furthermore, we study metrics based
on repated annotations of a training set. We define a way of assessing
the correlation between landmarks contrary to landmark coordinates. Finally,
we apply the proposed methods to a 2D data set consisting of outlines of
lungs and a 3D/(4D) data set consisting of sets of
mandible surfaces. In the latter
case the end goal is to construct a model for growth prediction and simulation.
Keywordsmaximum autocorrelation factors, maximum noise fractions, shape analysis, growth modelling
TypeConference paper [With referee]
ConferenceMedical Image Computing and Computer-Assisted Intervention - MICCAI 2002, 5th Int. Conference, Tokyo, Japan
SeriesLecture Notes in Computer Science
Electronic version(s)[pdf]
BibTeX data [bibtex]
IMM Group(s)Image Analysis & Computer Graphics