Title: Task Driven Perceptual Organization for Extraction of Rooftop
       Polygons
Authors: Christopher Jaynes, Frank Stolle, and Robert Collins
Affiliation: Computer Vision Research Laboratory, Dept. of Computer
             Science, University of Massachusetts, Box 34610, Amherst,
             MA 01003-4610
Abstract:
A new method for extracting planar polygonal rooftops in monocular
aerial imagery is proposed.  Through bottom-up and top-down construction
of perceptual groups, polygons in a single aerial image
can be robustly extracted.
Orthogonal corners and lines are extracted and hierarchically related using
perceptual grouping techniques.
Top-down feature verification is used so that features, and links between 
the features, are verified with local information in the image and weighed 
in a graph structure according to the underlying support for each feature.
Cycles in the graph correspond to possible building rooftop hypotheses.
Virtual features are hypothesized for the perceptual completion of partial 
rooftops.
Extraction of the 'best' grouping of features into  a building rooftop 
hypothesis is posed as a graph search problem.
The maximally weighted, independent set of cycles in the graph is extracted 
as the final set of roof boundaries.
Keywords: aerial imagery, building detection, perceptual grouping, graph search