We compare different unsupervised approaches to quantify the width of elongated structures in biological images. Two approaches detect first the centerline of the structure and rely on topological asymptotics of various cost functions and a fourth order elliptic PDE. The other approaches aim at detecting directly the boundaries of the structures: the gold standard is Canny's edge detector, and the last is based on topological asymptotics of a second order elliptic PDE.

This presentation is part of Contributed Presentation “CP6 - Contributed session 6”

Authors:

Jerome Fehrenbach (University of Toulouse)

Frederic de Gournay (INSA Toulouse)

Keywords:

image segmentation, partial differential equation models