TY - JOUR
T1 - A Geometric Framework for Stochastic Shape Analysis
AU - Arnaudon, Alexis
AU - Holm, Darryl D.
AU - Sommer, Stefan
PY - 2019/6/15
Y1 - 2019/6/15
N2 - We introduce a stochastic model of diffeomorphisms, whose action on a variety of data types descends to stochastic evolution of shapes, images and landmarks. The stochasticity is introduced in the vector field which transports the data in the large deformation diffeomorphic metric mapping framework for shape analysis and image registration. The stochasticity thereby models errors or uncertainties of the flow in following the prescribed deformation velocity. The approach is illustrated in the example of finite-dimensional landmark manifolds, whose stochastic evolution is studied both via the Fokker–Planck equation and by numerical simulations. We derive two approaches for inferring parameters of the stochastic model from landmark configurations observed at discrete time points. The first of the two approaches matches moments of the Fokker–Planck equation to sample moments of the data, while the second approach employs an expectation-maximization based algorithm using a Monte Carlo bridge sampling scheme to optimise the data likelihood. We derive and numerically test the ability of the two approaches to infer the spatial correlation length of the underlying noise.
AB - We introduce a stochastic model of diffeomorphisms, whose action on a variety of data types descends to stochastic evolution of shapes, images and landmarks. The stochasticity is introduced in the vector field which transports the data in the large deformation diffeomorphic metric mapping framework for shape analysis and image registration. The stochasticity thereby models errors or uncertainties of the flow in following the prescribed deformation velocity. The approach is illustrated in the example of finite-dimensional landmark manifolds, whose stochastic evolution is studied both via the Fokker–Planck equation and by numerical simulations. We derive two approaches for inferring parameters of the stochastic model from landmark configurations observed at discrete time points. The first of the two approaches matches moments of the Fokker–Planck equation to sample moments of the data, while the second approach employs an expectation-maximization based algorithm using a Monte Carlo bridge sampling scheme to optimise the data likelihood. We derive and numerically test the ability of the two approaches to infer the spatial correlation length of the underlying noise.
KW - Shape analysis
KW - Stochastic flows of diffeomorphisms
KW - Stochastic geometric mechanics
KW - Stochastic landmark dynamics
UR - http://www.scopus.com/inward/record.url?scp=85048455538&partnerID=8YFLogxK
U2 - 10.1007/s10208-018-9394-z
DO - 10.1007/s10208-018-9394-z
M3 - Journal article
AN - SCOPUS:85048455538
SN - 1615-3375
VL - 19
SP - 653
EP - 701
JO - Foundations of Computational Mathematics
JF - Foundations of Computational Mathematics
IS - 3
ER -