Title :
A Constant Average Time Algorithm to Allow Insertions in the LAESA Fast Nearest Neighbour Search Index
Author :
Mico, Luisa ; Oncina, Jose
Author_Institution :
Dept. Lenguajes y Sist. Informticos, Univ. de Alicante, Alicante, Spain
Abstract :
Nearest Neighbour search is a widely used technique in Pattern Recognition. In order to speed up the search many indexing techniques have been proposed. However, most of the proposed techniques are static, that is, once the index is built the incorporation of new data is not possible unless a costly rebuilt of the index is performed. The main effect is that changes in the environment are very costly to be taken into account. In this work, we propose a technique to allow the insertion of elements in the LAESA index. The resulting index is exactly the same as the one that would be obtained by building it from scratch. In this paper we also obtain an upper bound for its expected running time. Surprisingly, this bound is independent of the database size.
Keywords :
pattern recognition; LAESA index; constant average time algorithm; index element insertion; nearest neighbour search index; pattern recognition; Buildings; Heuristic algorithms; Indexing; Pattern recognition; Upper bound; distance; index; insertion; metric spaces; nearest neighbour;
Conference_Titel :
Pattern Recognition (ICPR), 2010 20th International Conference on
Conference_Location :
Istanbul
Print_ISBN :
978-1-4244-7542-1
DOI :
10.1109/ICPR.2010.951