Shape Features Extraction Using a Partial Differential Equation

06/13/2018

∙

by   Takayuki Yamada, et al.

∙

0

∙

share

This paper presents a unified method for extraction of geometrical shape features in binary image data using a linear partial differential equation (PDE). The PDE and functions are formulated to extract geometrical shape features, which are thickness, shape orientation and skeleton, all at once. The main advantages of the proposed method are it is free of any computation with respect to distance, it has no topological constraint of target image data and surfaces do not have to be distinguished to inside or outside. A one dimensional analytical solution is provided to validate the proposed method. Additionally, two- and three-dimensional numerical examples are shown to confirm the validity and usefulness of the proposed method.