Presentation 2006-03-16
An Accelerated High Dimensional Nearest Neighbor Search based on An Efficient Distance Computational Strategy
Koji Takemoto, Takekazu Kato, Toshikazu Wada,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose an accelerated Nearest Neighbor (NN) search algorithm in high dimensional space. The methods proposed so far can be classified into two types: 1) NN candidate narrowing and 2) pruning of distance computation. In high dimensional space over 30D, while NN candidate narrowing becomes brute force search, the latter method, pruning of distance computation is still effective for acceleration. For realizing more efficient NN search in high dimensional space, we integrate these two methods. As well, orthogonal expansion of patterns ordered by contribution ratio is incorporated for efficient pruning, because the distance computation starting from the most contributed component provides good approximation of the true distance. We confirmed through extensive experiments that our method is faster than existing methods.
Keyword(in Japanese) (See Japanese page)
Keyword(in English)
Paper # PRMU2005-240
Date of Issue

Conference Information
Committee PRMU
Conference Date 2006/3/9(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An Accelerated High Dimensional Nearest Neighbor Search based on An Efficient Distance Computational Strategy
Sub Title (in English)
Keyword(1)
1st Author's Name Koji Takemoto
1st Author's Affiliation Graduate School of System Engineering, Wakayama University()
2nd Author's Name Takekazu Kato
2nd Author's Affiliation Graduate School of System Engineering, Wakayama University
3rd Author's Name Toshikazu Wada
3rd Author's Affiliation Graduate School of System Engineering, Wakayama University
Date 2006-03-16
Paper # PRMU2005-240
Volume (vol) vol.105
Number (no) 673
Page pp.pp.-
#Pages 8
Date of Issue