Understanding the spatial organization of regions in images is a
crucial task, essential to many domains of computer vision. The
histogram of forces--a quantitative representation of the relative
position between two objects--constitutes a powerful tool dedicated to
this task. It encapsulates structural information about the objects as
well as information about their spatial relationships. Moreover, it
offers solid theoretical guarantees and nice geometric properties.
Numerous applications have been studied, and new applications continue
to be explored. For instance, force histograms can be compared through
similarity measures for fuzzy scene matching. They can be used for
describing relative positions in terms of spatial relationships modeled
by fuzzy relations. They can also be used for scene description, where
relative positions are represented by linguistic expressions. This
chapter reviews and classifies work on the histogram of forces. It
touches topics as varied as human-robot communication and spatial
indexing mechanisms for medical image databases.
|