Constraint relaxing on topological operators
in pictorial queries with null answer
 

Arianna D’Ulizia, Fernando Ferri, Patrizia Grifoni
IRPPS-CNR, via Nizza 128, 00198 Roma, Italy
arianna.dulizia, fernando.ferri, patrizia.grifoni@irpps.cnr.it

and

Maurizio Rafanelli
IASI-CNR, viale Manzoni 30, 00185 Roma, Italy
rafanelli@iasi.cnr.it

 
Abstract
 
This paper proposes an approach for relaxing constraints of a pictorial query when the answer returns a null value. In particular this approach, similarly with the 9-intersection matrix between two sets point, proposes to extend the representation of the symbolic graphical objects (SGO) in terms of interior, boundary, and complement points, applied to any configuration of two SGO. The suggested method considers three matrices: the first for representing the points cardinality of the intersection of interior, boundary and exterior of a polyline with the analogous three intersections of a polygon, the second for representing the polylines cardinality of the same intersection sets and the third for representing the polygons cardinality of the same intersection sets. Moreover, in order to relax topological constraints of any configuration, conceptual neighbourhood graphs between topological operators valid for the different pairs of SGO are proposed, as well as the connected complex matrices which enlarge the above mentioned 9-intersection matrix. A computational model for determining relations that are conceptually more similar for each configuration is also proposed and the polyline-polygon relation is studied.
 
Keywords:Human-computer interaction, Advanced query languages, Geographical databases, Constraint relaxing, Pictorial queries.