FitEllipsoid: a fast supervised ellipsoid segmentation plugin

Jérôme Fehrenbach (1, 2) , Bastien Kovac (1) , Pierre Weiss (1, 2)

1 ITAV - Institut des Technologies Avancées en sciences du Vivant
2 IMT - Institut de Mathématiques de Toulouse UMR5219

We propose a supervised segmentation algorithm dedicated to fitting ellipsoids in 3D images with a minimal amount of user inputs. Its computational core is based on an original approach to fit ellipsoids to point clouds in an affine invariant manner. The algorithm estimates the fitting ellipsoid in a small number of cheap iterations. The code is distributed as an open-source plugin to be used within the image analysis software Icy. Computational experiments show that collections of nuclei in fluorescence microscopy images can be segmented much faster than with more traditional 2D delineation approaches.

Mots clés: ellipsoid fitting; Supervised segmentation; open-source software; Icy plugin; fast convex optimization

Domaines: Traitement des images [eess.IV]; Optimisation et contrôle [math.OC]

Jérôme Fehrenbach, Bastien Kovac, Pierre Weiss. FitEllipsoid: a fast supervised ellipsoid segmentation plugin. 2017. ⟨hal-01520337v1⟩