This paper presents a boundary-based, topological shape de-scriptor: the distance profile. It is inspired by the LBP (= local binarypattern) scale space - a topological shape descriptor computed by a fil-tration with concentric circles around a reference point. For rigid objects,the distance profile is computed by the Euclidean distance of each bound-ary pixel to a reference point. A geodesic distance profile is proposed forarticulated or deformable shapes: the distance is measured by a combina-tion of the Euclidean distance of each boundary pixel to the nearest pixelof the shape´s medial axis and the geodesic distance along the shape´smedial axis to the reference point. In contrast to the LBP scale space, itis invariant to deformations and articulations and the persistence of theextrema in the profiles allows pruning of spurious branches (i.e. robust-ness against noise on the boundary). The distance profiles are applicableto any shape, but the geodesic distance profile is especially well-suitedfor articulated or deformable objects (e.g.applications in biology).


Janusch, I., Artner, N., & Kropatsch, W. (2017). Euclidean and Geodesic Distance Profiles. In International Conference on Discrete Geometry for Computer Imagery DGCI 2017: Discrete Geometry for Computer Imagery (pp. 307–318). Springer International Publishing. http://hdl.handle.net/20.500.12708/57088