Winner-Update Algorithm for Nearest Neighbor Search

This paper presents an algorithm, called the winnerupdate algorithm, for accelerating the nearest neighbor search. By constructing a hierarchical structure ,for each feature point in the 1, metric space, this algorithm can save a large amount of computation at the expense of moderate preprocessingand twice the memorystorage. Givena queiy point,the costfor computing the distancesfrom thispoint to all the samplepoints can be reduced by using a lower bound list of the distance establishedfrom Minkowski's inequality. Our experiments have shown that the proposed algorithm cansave a largeamountof computation,especiallywhen the distance between the query point and its nearest neighbor is relatively small. With slight modification, the winnerupdatealgorithm can also speed up the searchfor k nearest neighbors, neighbors within a specified distance threshold, and neighbors close to the nearest neighbor.

• Type: Conference
• Year: 2000
• Where: ICPR
• Authors: Yong-Sheng Chen, Yi-Ping Hung, Chiou-Shann Fuh
• Keywords: Computer Vision, Distance Establishedfrom Minkowski, ICPR 2000, Nearest Neighbor, Nearest Neighbor Search