Institutional Repository [SANDBOX]
Technical University of Crete
EN  |  EL

Search

Browse

My Space

Indexing images with multiple regions

Petrakis Evripidis

Full record


URI: http://purl.tuc.gr/dl/dias/0F7A012D-3D8D-4EDB-AB25-81CA7CD93E5F
Year 2001
Type of Item Conference Paper Abstract
License
Details
Bibliographic Citation Giannis Varelas, Paraskevi Raftopoulou, Euripides G.M. Petrakis, Evangelos E. Milios. ( 2005, Sep. ). Semantic Similarity Methods in WordNet and their Application to Information Retrieval on the Web. Presented at 7th ACM International Workshop on Web Information and Data Management (WIDM 2005). [Online]. Available:http://www.intelligence.tuc.gr/~petrakis/publications/petr01a.pdf
Appears in Collections

Summary

Similarity indexing using Spatial Access Methods (SAMs) like e.g., R-trees, assumesthat each data entity (or query) is represented by exactly one multidimensional point.However, for several applications, including indexing and retrieval of multimedia data like onedimensional signals and images, it is required that each data entity is represented by multiple pointsin a multidimensional space. This work extendsthe existing framework of indexing using SAMs to handle such data entities and has many desirable properties: For example, it provides index support for the two most common types of similarity queries, namely range and nearest neighbor (NN) queries, it returns exactly the same answers with the sequential scan method(only much faster) and, it works with any SAM and any data type. The effectiveness of the proposed approach is demonstrated using images with multiple regions.

Services

Statistics