IMAGE COMPLETION USING A DIFFUSION DRIVEN MEAN CURVATURE FLOW IN A SUB-RIEMANNIAN SPACE

Abstract

Dipartimento di Elettronica, Informatica e Sistemistica, Universita di Bologna, Bologna, Italy`citti@dm.unibo.itKeywords: Perceptual completion.Abstract: In this paper we present an implementation of a perceptual completion model performed in the three dimen-sional space of position and orientation of level lines of an image. We show that the space is equipped witha natural subriemannian metric. This model allows to perform disocclusion representing both the occludingand occluded objects simultaneously in the space. The completion is accomplished by computing minimalsurfaces with respect to the non Euclidean metric of the space. The minimality is achieved via diffusion drivenmean curvature flow. Results are presented in a number of cognitive relevant cases.