Showing papers by "Nikos Mamoulis published in 1998"

Algorithms for Querying by Spatial Structure

Abstract: Structural queries constitute a special form of content-based retrieval where the user specifies a set of spatial constraints among query variables and asks for all configurations of actual objects that (totally or partially) match these constraints. Processing such queries can be thought of as a general form of spatial joins, i.e., instead of pairs, the result consists of n-tuples of objects, where n is the number of query variables. In this paper we describe a flexible framework which permits the representation of configurations in different resolution levels and supports the automatic derivation of similarity measures. We subsequently propose three algorithms for structural query processing which integrate constraint satisfaction with spatial indexing (R-trees). For each algorithm we apply several optimization techniques and experimentally evaluate performance using real data.

Assessing multimedia similarity: a framework for structure and motion

Abstract: In this paper we address the issue of structural multimedia similarity, which is based on the relations between the individual objects that comprise a multimedia document. We propose a binary string encoding for 1D relations which permits the automatic derivation of similarity measures. We then extend it to various resolution levels and many dimensions and show that reasoning on spatiotemporal structure is significantly facilitated in the new framework, by applying it to multimedia presentation and motion similarity.

Image similarity retrieval by spatial constraints

Abstract: 1. ABSTRACT This paper deals with queries involving the retrieval of images that contain certain object configurations. Consider, for instance, that a user wants to “find all images where there exists a building adjacent to the west side of a park which is southwest and near a commercial center”. This query can be formulated as a constraint satisfaction problem (CSP) where the query variables are nodes of the corresponding constraint network and the image objects constitute the domain of each variable. The arcs of the network correspond to spatial constraints (e.g., adjacent ∧ west (X1,X2), southwest ∧ near (X2,X3)). Problems of the above nature are, in general, intractable. In addition, spatial constraints (e.g., southwest, near) lack universally accepted semantics and cannot always be modeled by crisp relations; a fact that further complicates query processing. This paper focuses on the development of effective methods that take advantage of the special structure of the spatial domain to achieve good average performance even for large images and queries. 1.1

Constraint-based algorithms for computing clique intersection joins

Abstract: 1. ABSTRACT Spatial joins constitute one of the most active research topics in spatial query processing. This paper deals with the processing of clique intersection joins using R-trees. A clique intersection join will retrieve all n-tuples of objects that pair-wise overlap. The corresponding MBR-based filter step retrieves n-tuples of rectangles that intersect at some common point. Here we modify three algorithms, first proposed in [13], for the specific problem and experimentally evaluate their performance using data sets of various densities. 1.1