A Hierarchical Geodesic Model for Longitudinal Analysis on Manifolds

  title={A Hierarchical Geodesic Model for Longitudinal Analysis on Manifolds},
  author={Esfandiar Nava-Yazdani and Hans-Christian Hege and Christoph von Tycowicz},
  journal={Journal of Mathematical Imaging and Vision},
  pages={395 - 407}
In many applications, geodesic hierarchical models are adequate for the study of temporal observations. We employ such a model derived for manifold-valued data to Kendall’s shape space. In particular, instead of the Sasaki metric, we adapt a functional-based metric, which increases the computational efficiency and does not require the implementation of the curvature tensor. We propose the corresponding variational time discretization of geodesics and employ the approach for longitudinal… 

Predicting Shape Development: a Riemannian Method

A novel prediction method that encodes the whole shape in a Riemannian shape space and outperforms deep learning-supported variants as well as state-of-the-art when applied to predict the shape of the right hippocampus under Alzheimer's disease and to human body motion.

A Soft-Correspondence Approach to Shape-based Disease Grading with Graph Convolutional Networks

This work presents a graph-based learning approach for morphometric classification of disease states that is based on a generalized notion of shape correspondences in terms of functional maps and shows that the approach can improve over state-of-the-art from geometric deep learning.



Sasaki metrics for analysis of longitudinal data on manifolds

The ability of these methods to distinguish differences in shape changes in a comparison of longitudinal corpus callosum data in subjects with dementia versus healthily aging controls is demonstrated.

Geodesic Analysis in Kendall’s Shape Space with Epidemiological Applications

This work analytically determines Jacobi fields and parallel transports and compute geodesic regression in Kendall’s shape space and can fully leverage the geometry via Riemannian optimization and thereby reduce the computational expense by several orders of magnitude over common, nonlinear constrained approaches.

A Nonlinear Hierarchical Model for Longitudinal Data on Manifolds

Large longitudinal studies provide information that is particularly valuable in medical studies. A problem that must be solved in order to realize the full potential is the correlation between

Geodesic B-Score for Improved Assessment of Knee Osteoarthritis

This work derives a consistent generalization of the recently proposed B-score to Riemannian shape spaces and exhibits improved discrimination ability over its Euclidean counterpart, which is demonstrated for predictive validity on assessing risks of total knee replacement.

Bi-invariant Two-Sample Tests in Lie Groups for Shape Analysis - Data from the Alzheimer's Disease Neuroimaging Initiative

Generalizations of the Hotelling's $T^2$ statistic and the Bhattacharayya distance for data taking values in Lie groups are proposed and validated in group tests revealing significant differences in hippocampal shape between individuals with mild cognitive impairment and normal controls.

Learning the spatiotemporal variability in longitudinal shape data sets

A generative statistical model to learn the spatiotemporal variability in longitudinal shape data sets, which contain repeated observations of a set of objects or individuals over time, is proposed and applied to the modeling of the progressive atrophy of the hippocampus in Alzheimer’s disease patients.

Hierarchical Multi-geodesic Model for Longitudinal Analysis of Temporal Trajectories of Anatomical Shape and Covariates

To the first longitudinal framework to model anatomical developments over time as functions of both continuous and categorical covariates on a nonlinear shape space, a new hierarchical multi-geodesic model is proposed.

Towards Shape-Based Knee Osteoarthritis Classification Using Graph Convolutional Networks

A transductive learning approach for morphometric osteophyte grading based on geometric deep learning using a combination of an intrinsic dimension reduction together with a graph convolutional neural network for high-dimensionality and non-Euclidean structure of shape space.