Biomathematics Seminar: Shape modelling using contours, phase fields, and binary fields, with applications to image segmentation
11 February 2011 15:15 in CM107 (Mathematical Sciences)
The study and modelling of shape has a long history, in biology,
statistics, and more recently, in computer vision and image processing.
For the most part, shape modelling has focused on families of shapes
consisting of perturbations of a given reference shape with a simple
topology. There are applications, however, where the family of shapes
involved does not have such a constrained behaviour. Cases where the
number of individual objects is unknown a priori, or where the topology of
the shape may be otherwise complex (for example network shapes), require
new techniques. The introduction of explicit long-range interactions
between shape boundary points allows these more complex cases to be
addressed, thereby creating a general shape modelling framework.
I will present several models developed within this framework, in
particular models of network shapes and of a 'gas of circles', and
describe how the transformation of models between different shape
representations (boundaries, phase fields, binary fields) allows the
analytical and algorithmic advantages of each to be exploited.
The models can be applied to the segmentation of entities from image data.
I will describe applications to the segmentation of road networks and tree
crowns from satellite and aerial images, and preliminary results on