The fshape framework for the variability analysis of functional shapes (Q2397743)

The subject of this paper are fshapes or functional shapes. In this framework, a geometrical shape \(X\), which is basically a submanifold, has ``associated with it a scalar field \(f: X\rightarrow R\) attached to every geometrical point of the support \(X\).'' Based on this point of view, the authors introduce ``a well-defined and well-structured ensemble \(F\) of fshapes''. Using fshapes, the authors establish well-posed variational problems where they can prove the existence of a minimizer. This is useful since it provides an existence result for an average shape from a population of noisy geometrical shapes. The authors also present a numerical implementation and numerical results that illustrate the methodology. In particular, they present an example with a population of retinal membranes with thickness maps.